Science topic

Survey Analysis - Science topic

Explore the latest questions and answers in Survey Analysis, and find Survey Analysis experts.
Questions related to Survey Analysis
  • asked a question related to Survey Analysis
Question
9 answers
If I ask to respondent a question and s/he has to answer the question from his/her perception then there is always a probability to answer the question in positive manner that do not underestimate the respondents. As for example, if I ask a question: Do you think you have the following capacity of your own?
Question: I can manage more than one complexity in a time.
and the scale for this question: Agree, no opinion, disagree. 
Then every one will try to highlight the positive site and the answer will be yes, I agree.
In that case is there any way to avoid the error or any methodology to find out the real answer?
Please give me your valuable suggestions.
Thank you. in advance for your anticipated cooperation.
Relevant answer
Answer
Have a look at the source Chuck A Arize didn’t mention: https://aytm.com/blog/author-bias-how-to-avoid-asking-leading-questions/ The exact phrases/sentences “Be conscious of scenarios where you may be asking leading questions and understand the resolve by; using neutral language, refrain from embellishing questions, and avoiding absolute words. Applying these tools will allow you to extrapolate authentic and meaningful data from your respondents.” can be found if you scroll down under the heading “Practice Makes Perfect – The Takeaway”.
Best regards.
  • asked a question related to Survey Analysis
Question
11 answers
I have conducted a survey on paper. The questionnaire consists of 82 items divided into 15 subscales. That is, 5-6 items belong to a scale/variable I want to explore. I use a 6 point Likert scale for each of the items. Even though the questionnaire clearly states that the respondents should choose only one response per item, some respondents have chosen two options (e.g. 1 and 2, or 3 and 4). How do I handle this? I would really appreciate suggestions, thanks!
Relevant answer
Answer
Perhaps the safest strategy would be to treat these responses as missing values and subsequently use multiple imputation to address missing item scores. Another strategy could be to assign a score of 1.5, 2.5, etc. in these cases if you think it is plausible that these participants were trying to respond "in between" 1 and 2 (or 2 and 3 etc.).
  • asked a question related to Survey Analysis
Question
5 answers
Hello everyone,
Could you recommend courses, papers, books or websites about mixed survey analysis methodologies?
Thank you for your attention and valuable support.
Regards,
Cecilia-Irene Loeza-Mejía
Relevant answer
Answer
The following RG link is also very useful:
  • asked a question related to Survey Analysis
Question
46 answers
Dear researchers,
I have estimated internal consistency for a questionnaire with 50 items (five -likert point) by alpha cronbach. Alpha was .98.
What is your interpretation?
What can the cause be from your point of view?   
Thanks for your help.
Relevant answer
Answer
Exactly, Cronbach's alpha indicates whether the items measure the same construct. The minimum acceptable value for Cronbach's alpha ca 0.70; Below this value the internal consistency of the common range is low. Meanwhile, the maximum expected value is 0.90; Above this value is perceived as redundancy or duplication.
  • asked a question related to Survey Analysis
Question
1 answer
Hi everyone! I am currently conducting research on how the U.S. federal government’s response to COVID-19 shaped its perceived legitimacy among Americans, and is this mediated by affective polarization. My variables are:
IV: Party identification (“Generally speaking, do you think of yourself as a Democrat, a Republican, or Independent?” A follow up question was used: “Which of the following parties do you identify with the most? Strong Democrat, Not very strong Democrat, Strong Republican, Not very strong Republican, Independent, Independent - Dependent, Independent - Republican.”)
DV: Legitimacy (Six scale items were measured across a 5-point Likert scale ranging from Strongly Agree to Strongly Disagree)
Mediator: Affective Polarization (8-point trait item scale to measure positive and negative attitudes towards a respondent’s preferred political party compared to their opposed one - Democrat and Republican: "delighted,” “angry,” “happy,” “annoyed,” “joy,” “hateful,” “relaxed” and “disgusted.” A Feeling Thermometer was also used).
This is my first survey analysis and my first time using R, and needless to say there's been a steep learning curve on teaching myself how to use the software.
Question: does anyone have a good idea on where to start to tackle this? Do I start with CFA, re-measure Cronbach's Alpha for the scales? How do I aggregate the multi-items scale items? Is it necessary for me to use or discard one measure of the other? You're help is very much appreciated :-)
Relevant answer
Answer
Hi Lisa, I see this is old, but this may help you or others.
One of the most popular packages for mediation and moderation is PROCESS by Hayes - who has written a number of texts and articles on these stats and developed software for SPSS and SAS. There is now an R package here: https://www.processmacro.org/download.html
There is a nice 'how to' for PROCESS in R, here: http://www.regorz-statistik.de/en/mediation_process_for_r.html
The mediation package is also popular with many useful options: https://www.rdocumentation.org/packages/mediation/versions/4.5.0
And, here is a nice page on how to do mediation in R: https://www.r-bloggers.com/2019/08/how-to-do-mediation-scientifically/
  • asked a question related to Survey Analysis
Question
9 answers
Hi everyone,
I would like to run a Kruskal-Wallis test on a large sample with a likert scale. I am having difficulties figuring out how to process with the ''rank_avg'', and rank sum with such a large sample. Here's how my data is presented for a typical question:
-Marketing department, 1200 respondants, 88% are favorable, 4% neutral and 8% not favorable to the question
-Eng. department, 500 respondants, 77% are favorable, 3% neutral, 20%... and so on
I would like to assess if results from each department are statistically significantly different from one another. Maybe this is not the right test to use, I would like your help since it is my first survey analysis.
Thank you and best regards
Relevant answer
Answer
Hello all,
In response to the quotation furnished from the reference given by Luiza Betina Petroll Rodrigues , I think the quote is, at best, misunderstood or at worst, just poor advice.
1. It is true that the hypotheses tested by non-parametric tests are different from those evaluated by parametric tests, a point that some researchers may fail to appreciate. For example, compare the independent t-test with the Mann-Whitney (or Wilcoxon-Mann-Whitney) test. The independent t evaluates the null hypothesis that two samples are drawn from populations having identical means (e.g., mu1 = mu2). The Mann-Whitney test evaluates the null hypothesis that two distributions are identical, versus the alternative that one is displaced above the other. Only if you assume symmetry of distributions does the test become an evaluation of equality of medians.
2. Sample size does not change the applicability of the choice between parametric, non-parametric, and exact/permutation tests. Why not? (a) If data are not of interval/ratio strength, then parametric tests (and simple statistics such as a mean) are not meaningful; (b) Textbooks dating back to pre-computer times would often declare N = 30 as a "large sample size," obviating the need for non-parametric methods. However, this was chiefly an argument which rested on the central limit theorem. Work by Theodore Micceri: The unicorn, the normal curve, and other mythical creatures, and Rand Wilcox suggest that normality may be an unrealistic expectation in the first place, and that the degree of departure from normal may well require far larger N to yield good fit in the tails of the distribution of sampling means; (c) Tests like the Mann-Whitney require an assumption of homogeneity of variance, which is often overlooked.
3. There are both exact/permutation tests and, the approximations thereto (bootstrap or resampling), which can change the hypotheses, but give good insight as to what's going on without assumptions of distribution shape or homogeneity.
A belated happy new year to all.
  • asked a question related to Survey Analysis
Question
6 answers
Do you have one or more reference resources that you can recommend for the necessity of reverse coding/scoring in surveys?
Relevant answer
Answer
Some researchers use negatively worded items on a survey instrument to contain acquiescence bias; however, the effect of such a strategy on response accuracy and instrument validity could preponderate its benefits. In other words, mixing positively and negatively worded items on a survey instrument might beget threats to its validity and reliability. The following could be of interest.
Chyung, S. Y. (Yonnie), Barkin, J. R., & Shamsy, J. A. (2018). Evidence-based survey design: The use of negatively worded items in surveys. Performance Improvement, 57(3), 16–25. https://doi.org/10.1002/pfi.21749
Suárez-Álvarez, J., Pedrosa, I., & Lozano, L. M. (2018). Using reversed items in Likert scales: A questionable practice. Psicothema, 30.2, 149–158. https://doi.org/10.7334/psicothema2018.33
Weijters, B., & Baumgartner, H. (2012). Misresponse to reversed and negated items in surveys: A review. Journal of Marketing Research, 49(5), 737–747. https://doi.org/10.1509/jmr.11.0368
Good luck,
  • asked a question related to Survey Analysis
Question
3 answers
The initial use of meta anysis is to compare results from clinical trials.
However, could I use it for data obtained from questionaire.
For instance, the question for dentist would be: ''Do you sterilize your devices regularly?''. I will try to combine responsed from different studies.
Is there any manual or framework on that methodology?
Thanks,
Aleksandar
Relevant answer
Answer
If surveys provide binary answers (yes/no), you can perform a meta-analysis of single proportions using the inverse variance method (with Freeman–Tukey double-arcsine transformation) or a generalized linear mixed model. It is easy to implement in Stata or R, eg. with Metaprop.
  • asked a question related to Survey Analysis
Question
8 answers
Hello,
I have been using the Thomas Lumley's "survey" package for complex survey analysis in R. I understood that multinomial regression model is not developed yet in "survey" package. Is there any other solution to address this problem?
Best wishes,
Uuree.
Relevant answer
Answer
For anyone just joining this discussion, there is a new package svyVGAM that is able to fit multinomial logistic regression models with a complex survey design.
There is also svyolr for those interested in ordered logistic regressions.
For marginal effects, it seems that Stata (svy: ologit or mlogit) offer wider options.
Hope this is helpful
Sebastián
  • asked a question related to Survey Analysis
Question
2 answers
Through the articles and video I've seen in the quantification of elements by XPS survey of Si based materials, I noticed that only Si 2p is quantified. Why Si 2s is ignored?
Thanks for any reply.
Relevant answer
Answer
You can quantify on the Si(2s) instead of the the Si(2p) if desired as the area will be adjusted for the orbital sensitivity factor.
However to relate exactly to your question, if you quantified the Si(2s) AND the Si(2p) then you would be broadly doubling the amount of Si you have in your sample - so getting an eroneous result. Its the same for any element - you only quantify one (usually the most intense) set of peaks
  • asked a question related to Survey Analysis
Question
7 answers
Hello!
So, here is the story. I was give this Likert scale data for analysis, and I just can't get it how I should deal with it. It is a 1-7 scale with answers ranging from 1 being "extremely worse" to 7 being "extremely better". But here is the problem, 4 is "same as was before" and questions introduce the changes as an effect of a different variable, which is work from home (for example, "Compared to work from office, how much has your ability to plan your work so that it was done on time changed when working at home?").
Questions are separated into some groups to form variables, and mean should probably show each person's opinion on the change, right? But it just seems too strange to me to work with just 1 parameters and not go through full comparison of now vs before as 2 different constructs.
If you have any works or insight on the topic, can you please help me?
All the best and take care!
Relevant answer
Answer
I agree with David L Morgan, Likert-scored items are ordinal. Don't worry about 4 (same as was before), it is a neutral option in Likert scale. Based on data distribution, you can use different statistical tests.
  • asked a question related to Survey Analysis
Question
4 answers
I am trying to perform the cell-weighting procedure on SPSS, but I am not familiar with how this is done. I understand cell-weighting in theory but I need to apply it through SPSS. Assume that I have the actual population distributions.
Relevant answer
Answer
I might be misunderstanding your question, or his answer, but in my reading of what you are trying to do, I think the approach suggested by David Morse is missing a final step.
I'll assume, as David did, that the population, with N=3200, consists of 500 cases (or 15.625%) in subgroup A, 700 (21.825%) in B, and 2000 (62.5%) in C. I'll assume, further, that you have a sample, with n=80, that includes 10 cases (or 12.5%) in A, 20 (25%) in B, and 50 (62.5%) in C. If so, then the cell weights for your sample should be (15.625/12.5 = 1.25) for cell A, (21.825/25 = 0.875) for cell B, and (62.5/62.5 = 1.0) for cell C. That will keep your weighted sample size at 80, the same as your unweighted sample size, but will make the proportion of cases in A, B, and C in your weighted sample equal to the population proportions.
Forming the weights as ratios of the % in the population divided by the % in the sample will inflate the under-represented cells and deflate the over-represented cells in your sample by exactly the right amount.
If, instead, you also want to make the total number of cases in your sample equal to the total population size, then each of the three initial weights (1.25, 0.875, and 1.0) should be multiplied by (3200/80 = 40), yielding three new weights (50, 35, and 40).
Multiplying by the ratio of population size divided by sample size inflates all of your initially weighted sample counts by exactly the right amount to equal the population count.
  • asked a question related to Survey Analysis
Question
3 answers
The question is actually very specific. I am doing a survey analysis.
I have 3 variables. Year of college education (4,5,6), whether the person has passed an exam (yes/no) and a score they got on a test. How can I compare if there is a difference in test scores only between 4th year students and people who passed or didn't pass the exam?
Relevant answer
Answer
Thanks guys
  • asked a question related to Survey Analysis
Question
13 answers
I hope to conduct a series of interviews/questionnaire surveys to collect information regarding urban flood management and the use of software tools for the same.
Fundamentally, decision-makers, flood modellers, general public and software modellers/developers are in my expected audience.
Could you please suggest what personal information should be considered when weighing them?
My assumptions are as follow;
1. Decision Makers: The age, level of education, years of service, the level in the organization, no of participations/decision makings in actual flood management activities
2. Flood modellers: educational status (MSc/PhD etc), years of experience, no of participations/decision makings in actual flood management activities
3. Software developers: years of experience, no of contributions in actual flood management software development and the role he/she played
4. General Public: The Age, the level of flood-affected to the person, educational level, experience with floods
Relevant answer
Answer
I appreciate the request to comment, Rmm, but I don't think I know enough about your particular problem domain.
That's the thing about applying weights to survey respondents - making responses from one person, or a group of people, more important than those of another. You would do this if you have a legitimate reason to think that one group is severely under-represented in your sampling frame, or in your final sample. Or if you have a theoretical reason for giving greater value to the responses of some, and lesser value to others.
You need to have a theory, and/or good evidence, to support the use of weights in the first place and some ideas about how much those weights should apply.
Thinking about it some more, the purpose of your research is likely to be important too. If you're interested only in the value of real estate affected by floods then your weights may apply to the value associated with the people/organisations you survey. If you are interested in the effects on people's homes then you may minimise commercial real estate and apply weights based on the sizes of families.
  • asked a question related to Survey Analysis
Question
4 answers
The Prospect Theory (Kahneman, D., & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2), 263-291) defines a "value function". I want to know if there is a way to estimate that curve based on answers of the questionnaires (prospects).
In the cumulative prospect theory they estimate some parameters but I didn't understand how (Tversky, A., & Kahneman, D. (1992). Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and uncertainty, 5(4), 297-323.).
Can anyone could make a more simple explanation for me?
Relevant answer
Thanks professor Paul Louangrath , but I still don't know how to formulate this regression using logistic function because we don't have explicit values of X and Y. For a given prospect ("choice") we have two alternatives, each one with a pair expected value/probability.
  • asked a question related to Survey Analysis
Question
10 answers
I have a question in my questionnaire regarding purchase intention, and the options to choose the answer are:
  • Definitely Not
  • Probably Not
  • Possibly
  • Probably
  • Definitely
From this question, I need to figure out the relation between purchase intention and three other factors (that are asked in 15 different likert scale questions)
I am doing a pretest, and so far 8 people have filled in the questionnaire and 7 of them have chosen 'Possibly' as their answers for the purchase intention question.
So, my question is if for example 90 percent of respondents in the final questionnaire choose the same answer for that question, can I still get the meaningful analysis from my data?
Relevant answer
Answer
Assuming that this is only a pretest and not the final application of the survey, you have just demonstrated the value of a pretest. The question, as currently designed, is not providing you with enough information to be useful in other types of analyses. One possibility is to increase the number of items to the scale (e.g., moving to a 10-point scale); people still would tend to give answers in the middle, but there hopefully would be more variation in their responses. Whether that variation would be meaningful, however, is another question.
It seems that people are picking the middle option because they don't know what they would do. (As a side note, I think "possibly" is probably not the best word choice. Based on the wording, it is a somewhat positive response, but as the middle of five items, I think people are treating it as a neutral response.) One thing you might investigate is whether they really don't have any strong feelings one way or the other, or they have conflicting feelings, leading to a overall undecided responses. Each topic could be pursued further, to determine why they don't care or what the conflicting feelings are. Maybe you might also want to investigate what new information or change in features would create the greatest change in their feelings. This might require focus groups and/or another pretest before you are ready to field the final instrument.
  • asked a question related to Survey Analysis
Question
3 answers
Hello,
This is fascinating work. As a Workforce specialist in this field, I'm curious if you've had the chance to conduct similar analyses on the NCI Staff Stability Survey. Perhaps on correlations between staff stability and quality outcomes? Thanks so much!
Best,
J. Haggerty
Relevant answer
Answer
To follow up on Val - yes, at the University of Minnesota's Institute on Community Integration we are currently analyzing the NCI Staff Stability data.
  • asked a question related to Survey Analysis
Question
12 answers
I am working on validating a questionnaire and I need to ensure that there are few (or no) outliers that might affect the factor analysis process. Is the outlier labeling technique (Hoaglin, Iglewicz) applicable to non-normal data?
Relevant answer
Answer
You can just use upper and lower quantiles. We use nonparametric statistical methods to analyze data that's not normally distributed. In the same way, instead of using standard deviation, you would use quantiles. That is, you can say to assign NaN to values greater than 95% and less than 5% of the data set.
  • asked a question related to Survey Analysis
Question
32 answers
The Slovin's Formula is quite popularly use in my country for determining the sample size for a survey research, especially in undergraduate thesis in education and social sciences, may be because it is easy to use and the computation is based almost solely on the population size. The Slovin's Formula is given as follows: n = N/(1+Ne2), where n is the sample size, N is the population size and e is the margin of error to be decided by the researcher. However, its misuse is now also a popular subject of research here in my country and students are usually discourage to use the formula even though the reasons behind are not clear enough to them. Perhaps it will helpful if we could know who really is Slovin and what were the bases of his formula.
Relevant answer
Answer
If you use statistical models (like Ttest, ANOVA, Pearson r, regression analysis, path analysis, SEM, among others) to test the hypotheses of your study, then I suggest you conduct a statistical power analysis in computing your minimum sample size. Sample size is a function of the following components: effect size, errors in decision (Type I and Type II), complexity of the statistical model, among others. Statistical power analysis or simply power analysis is finding the optimal combination of the said components. You can use the G*Power software which is downloadable for free. Just search it in google.
Slovin's formula has been taught by "irresponsible professors" in the Philippine Colleges and Universities. Sorry for the strong words, but that is true. Because they dont have formal training about Statistics, they taught the wrong things to their students. They are like blind people guiding another blinds.
Anyway, you may read a published article titled "On the misuse of Slovin's Formula".
I am on travel now. I am just using my phone to reply your message. I will email to you later the said article if you want. If interested, just email me at johnny.amora@gmail.com and then I will send you some materials including the said article.
  • asked a question related to Survey Analysis
Question
1 answer
As a part of my MTech Research Project, I am doing a Questionnaire Survey for analysis the Concurrent Delay Scenario in Indian Construction Industry. Expert opinion is required for the same. Please take out some of your valuable time and help me by filling up the following survey:
Regards,
Chintan Munvar
Relevant answer
Answer
Good project but include some case project also
  • asked a question related to Survey Analysis
Question
8 answers
I have conducted an online survey to assess public opinions to the Soft Drink Industry Levy (SDIL/ sugar tax) in the UK. The survey also investigates current soft drink consumption habits, and expected effectiveness of the SDIL. The survey also includes demographic questions such as age, gender, income and health.
I am now starting analysis. I have been advised to conducted one way ANOVA and logistic/ ordinal regressions in SPSS, however I am not sure which variables to use for which test, and how many participants/ data points are needed to conduct each test to make them statistically valid. Any help is much appreciated!
Relevant answer
Answer
You're welcome!
  • asked a question related to Survey Analysis
Question
7 answers
I am still collecting survey tools used by researchers who have used in studies of transgender men and women. The tools will be used to inform the development of a new survey tool that may be added to CDC's National HIV Behavioral Surveillance (NHBS) survey set. If you are willing to share your survey tool (and have not already done so) please do so.
All the best,
Stephen 
Relevant answer
Answer
I follow the question
  • asked a question related to Survey Analysis
Question
2 answers
I have been using R for quite a while now but came across data that requires the use of survey weights for proper analyses. Therefore I tried to use the "survey"-package which seemed to work fine at first (after some struggle - but nonetheless, it was okay) - but when double checking my work with STATA, I realized my numbers were WAY off.
Despite reading through the (sparse) documentations and examples there are, and even buying the package-author's book, I still don't feel very confident with this type of work. Nonetheless I would love to learn to master this package as survey analysis comes up quite often in my research, and I would love to continue working in R.
Any help and advice will be welcomed! Thank you!
Relevant answer
Answer
Dear Jochen
Please, use the following link I hope will be useful for you.
Good Luck
  • asked a question related to Survey Analysis
Question
5 answers
Im in the process of finalizing my dissertation, and im finding some difficulties to analyse my survey findings.
Relevant answer
Answer
Claire, this is a follow-up from my earlier input. It is your hypotheses or research questions that will determine the types of statistical analysis that will be used.
If you are dealing with relationships you will have to use regression analysis If you are dealing with correlation analysis, you use the appropriate one. If you're are making comparisons you will have to use analysis of variance. If you are doing classification, you have to use factor analysis or its like.
This is why both expertise and experience are necessary.
  • asked a question related to Survey Analysis
Question
19 answers
If I pick up data from a survey for only 10% and randomly generate the rest of 90% from an application. (based on the 10%)  Will this work?  I am in IS discipline.
I think many people do simulate things in other domains too.
Relevant answer
Answer
What is your purpose for creating simulated data? If your study is entirely methodological, then the source of the data may not matter. But if you are attempting to describe a population, or perform an experiment, then adding simulated data does not help you to accomplish either task, and may actually get in the way. You may be adding random error that attenuates the relationship you are looking for, or may be artificially creating the appearance of a relationship that does not actually exist.
  • asked a question related to Survey Analysis
Question
3 answers
I need to translate and understand the statistics. If you are online I need your help like now. Thank you
Relevant answer
Answer
check out the MAPI site. They are located in France. They are a library of measures from around the world. They also have translations. I hope this helps. 
  • asked a question related to Survey Analysis
Question
14 answers
A group of residents from two hospitals answered a survey anonymously scoring their confidence from 0 to 10 then a lecture provided. A follow up survey was answered by the same group using the same score. As I read, I can't use the paired t test as this requires pairing the same person which I can't tell due to the anonymity. Running independent t tests would make me lose power unnecessarily? 
I would appreciate the help..
Relevant answer
Answer
The paired t-test could only apply if you had some identity variable for matching so that you derive  a 'difference' variable. As it stands, you may have to use two independent samples t- test with pre and post test sessions as the groups. You may only need to document this as a caveat in the interpretation of the results. Using non parametric version of paired t test may not help a s much because even with that, you needed to pair up observations from the two two time points. But again, you could only think of non parametric equivalent  if you conduct some preliminary analysis and realize some skewness of concern. Your case stands aloof.
  • asked a question related to Survey Analysis
Question
3 answers
I've carried out a questionnaire, where participants had to rate on a likert scale 1-5 a list of response strategies with respect to their ability addressing a certain issue. In total there are five such major issues and they were asked to rank the strategies under each separately. What is the best analysis technique to analyse and summarize the above data.
Relevant answer
Answer
I agree with Muayyad Ahmad that you have 5 sub scales. the next question is whether or not they are strongly correlated with each other. If so, they could form a single scale.
Because there have been so many questions here about the scaling of Likert-scored items, I have compiled a set of resources on this topic:
  • asked a question related to Survey Analysis
Question
10 answers
My question was based on the knowledge, attitude, and behavior on Disaster Reduction. The respondent answer only (Agree, or Disagree or No idea).Results will be shown with comparative both of two group. I wanna know which statistics should I do for these data?
Relevant answer
Answer
Most of the answers are telling you to do the wrong thing!!
This is not an ordinal scale. Agree and disagree are opinions, while 'no idea' is no opinion. 
There are then two steps in the analysis. The first is
- what distinguishes between people who have an opinion and people who do not have an opinion?
and the second is
- among people who have an opinion what distinguishes between people who agree and people who do not.
Treating 'no opinion' as half way between agreeing and disagreeing is missing the point. If a person has no idea, it doesn't mean that they belong on a continuum between agreeing and disagreeing. To treat the scale as ordinal, the middle item would have to be "neither agree nor disagree" and (and this is important) the question should be one that everyone can understand. So if I ask you whether you agree that relations between China and Taiwan have deteriorated lately, you might not be able to agree or disagree, but you would understand the question. But if I asked you whether Mozart modelled his requiem on Handel's funeral music for queen Cariloline, are would agree or disagree provided you had a background in the musicology of the period, but for most people the question is inappropriate. Not having an opinion in this case is because the person could not reasonably be expected to have one. 
Does that make sense?
(By the way, I think there's a strong case that Mozart did know the Handel work, but it's not my specialist area)
  • asked a question related to Survey Analysis
Question
11 answers
1. What are African American male’s perception of mental illness?
2. How much does stigmas, associated with mental illnesses impact African American men understanding of mental illness?
There is a need for additional studies to identify the perceptions of mental illness within the black population, specifically with African American men. The researcher will use qualitative research techniques to explore the African American male participants' perceptions of mental illness and mental health services. The focus of this study will examine 25 African American men, in Tallahassee, Florida, ranging between the ages of 18 and above. The selection of the specific geographical area is a convenience sample.
The purpose of this study is to explore African American adult men perceptions of mental illness. The researcher will explore stigmas and misconceptions of mental illness among the participants. Additionally, the researcher will attempt to identify if such stigmas and misconceptions influence the participants willingness to seek mental health services. The use of qualitative research is to identify any patterns and gain general knowledge of the phenomen, based on participant’s opinions.
Relevant answer
Answer
Inn general, case studies would treat each individual as a case and make comparisons across those case. I suggest you look at Robert Yin's work as a guide.
In contrast, phenomenology would look at the in-depth, lived experience of each individual. There are any number of divisions within phenomenology, ranging from those that are heavily influenced by philosophers to those that are similar to content analysis (e.g., interpretive phenomenological analysis).
If you are not already familiar with the more philosophical versions of phenomenology, it would probably take considerable effort to master one of them.
Also, there are any number of other approaches to collecting and analyzing qualitative data, so you might say something more about why you have chosen these two.
  • asked a question related to Survey Analysis
Question
9 answers
I am an MA student in Dance Practises and I have to produce a research proposal poster for my final dissertation, I am going to 5 dance companies who teach disabled children to find out the methods they use and why (Physical guidance, tactile etc) I have looked at Social constructivist theory and using qualitative research methods, but would I use a mixed methods approach to analyse my information or stick with qualitative? Sorry its all very new to me!
Relevant answer
Answer
Hi, Julie. I recommend you do a content analysis - thematic analysis. The categorization process can be based on previously defined concepts (theory),  or you can build the category grid from the answers obtained in the questionnaires and interviews (empirical procedure)
  • asked a question related to Survey Analysis
Question
2 answers
I am currently conducting a dissertation for my final year at university. My research question is relating to attachment and giving behaviour. I have already tried the ECR-R, the 32-item list questionnaire, however, I am not achieving a high response rate due to the length.
I have heard about the ECR-S being only 12 items, and there is evidence that suggest that this questionnaire is just as a effective as the long 32-item ECR-R. Does anyone have any experience with the anaylising the data associated with the ECR-S attachment questionnaire? 
Relevant answer
Answer
Thank you very much! I have contacted the authors of the article to see if they could send me the scale and instructions of data entry, and analysis.
Thank you once again 
  • asked a question related to Survey Analysis
Question
5 answers
 i used to collect data by questionnaire, from  the same unified unit of analysis ( example; employees only), now i have a question, if i have to variables in my research model that cant be collected from the same unit of analysis ( example questionnaire for employees and other for  org. customer ) is it differ in analysis, or the same  analysis  techniques could be used?
Relevant answer
Answer
Hi, Dear Raghda A. Younis
It depends on your research questions and hypothesis if you want to analysis the influence or the role, then t-test, correlation matrix through factor analysis and regression are the right statistical analysis,
Best Regards, 
  • asked a question related to Survey Analysis
Question
3 answers
I am studying the effect of knowledge management on service quality, but I need to measure service quality by employees as the respondents of my survey how can I do that and I need recent studies on service quality for my literature review 
Relevant answer
Answer
Please also refer to the study by Bitner et al (1994), where the assessment/focus is placed on employees in relation to service quality, namely: from the employees’ point of view, what kinds of events lead to dis/satisfying service encounters? (p. 95). Similarly, see: di Mascio (2010); Gremler and Gwinner (2000); Bettencourt and Gwinner (1996); Mittal and Lassar (1996), etc. 
The following papers may be helpful for an understanding on service quality, and the role of service employees in service encounters.  
  • Bettencourt, L. A. and Gwinner, K. (1996) Customization of the service experience: the role of the frontline employee, International Journal of Service Industry Management, 7, 2, pp. 3-20.
  • Bitner, M. J., Booms, B. H. and Mohr, L. A. (1994) Critical service encounters: The employee's viewpoint, Journal of Marketing, 58, 4, pp. 95-106.
  • di Mascio, R. (2010) The Service Models of Frontline Employees, Journal of Marketing, 74, 4, pp. 63-80.
  • Farrell, A. M., Souchon, A. L. and Durden, G. R. (2001) Service Encounter Conceptualisation: Employees' Service Behaviours and Customers' Service Quality Perceptions, Journal of Marketing Management, 17, 5-6, pp. 1-14.
  • Gremler, D. D. and Gwinner, K., P. (2000) Customer-employee rapport in service relationships, Journal of Service Research, 3, 1, pp. 82-104.
  • Grönroos, C. (1984) A service quality model and its marketing implications, European Journal of Marketing, 18, 4, pp. 36-44.
  • Hansemark, O. C. and Albinsson, M. (2004) Customer satisfaction and retention: the experiences of individual employees, Managing Service Quality, 14, 1, pp. 40-57.
  • Mittal, B. and Lassar, W. M. (1996) The role of personalization in service encounters, Journal of Retailing, 72, 1, pp. 95-109.
  • Parasuraman, A., Zeithaml, V. A. and Berry, L. L. (1985) A conceptual model of service quality and its implications for future research, Journal of Marketing, 49, 4, pp. 41-50.
  • Zeithaml, V. A., Berry, L. L. and Parasuraman, A. (1996) The behavioral consequences of service quality, Journal of Marketing, 60, 2, pp. 31-46.
  • asked a question related to Survey Analysis
Question
6 answers
I administered a survey to parents of autistic children to see if a specific carbohydrate diet helped to improve the outcomes of sleep etc.I have around 6 such outcomes and the results are in form of yes/no. Also, there are demographic variables of income, ethnicity, sex. I am not sure what would be the best test to use. Also, any suggestions on how can I categorize data with demographic variables and 6-7 outcomes to test the intervention's success or failure? Thanks
Relevant answer
Answer
 Hi,
Can  Mc Nemar test be used for pre-post data if the categorical variable has more than two levels for example level of pain: None, Mild, Moderate, Severe? Thanks
  • asked a question related to Survey Analysis
Question
1 answer
I'm currently working on a systematic review of the impact of early palliative care intervention for patient diagnosed with advanced cancer. Selected publication that assessed the impact of those interventions on patient's quality of life have used different quality of life questionnaire. I already have found some MID value for QoL questionnaire but didn't ind nothing about the FACIT-Pal and FACIT-sp questionnaire.
So I would like to know if someone knows these values or some goods references about it.
Thanks
Relevant answer
Answer
The questionnaires are the same, they change government control
  • asked a question related to Survey Analysis
Question
7 answers
I applied a questionnaire to social workers, but the number of completed questionnaires is very low (9). On a item I obtain 4 for option a, 3 for option b and 3 for option c. Bootstrap can be used to estimate the distribution?
Relevant answer
Answer
Bootstrap won't present you the "true" distribution of you variable of interest, except rather an approximation that might be useful in calculating parameters of the true distribution.
The idea is extremely easy: you sample with replacement NN cases from your dataset of NN observations the same way as you sampled your data from the population. In R that would look like this
age <- c(21,81,85,27,39,61,15,20,39,40,87,87,69,59,54,71,66,88,1,2)
N <- 20
age_boot <- matrix(NA, 100, 20)
for (i in 1:100) {
  age_boot[i, ] <- sample(age, N, replace=TRUE)
}
or simpler but more "hacky" way:
age_boot <- replicate(100, sample(age, N, replace=TRUE))
  • asked a question related to Survey Analysis
Question
90 answers
Hi everyone. I am working on my quantitative chapter of my thesis and I would like to ask you about handling close ended questions using 5-point Likert scale questionnaire. My questionnaire is looking at students’ perspective towards a course called (Intensive English as a foreign language).
I have been looking at literature and I find it more confusing when it comes to cell range. I came across two methods of Mean distribution of the findings.
First method:
To determine the minimum and the maximum length of the 5-point Likert type scale, the range is calculated by (5 − 1 = 4) then divided by five as it is the greatest value of the scale (4 ÷ 5 = 0.80). Afterwards, number one which is the least value in the scale was added in order to identify the maximum of this cell. The length of the cells is determined below:
  • From 1 to 1.80 represents (strongly disagree).
  • From 1.81 until 2.60 represents (do not agree).
  • From 2.61 until 3.40 represents (true to some extent).
  • From 3:41 until 4:20 represents (agree).
  • From 4:21 until 5:00 represents (strongly agree).
Second method is the traditional way:
  •  mean score from 0.01 to 1.00 is (strongly disagree);
  •  to 2.00 is (disagree);
  • from 2.01 until 3.00 is (neutral);
  • 3.01 until 4:00 is (agree);
  • mean score from 4.01 until 5.00 is (strongly agree)
My questions are:
1             Which method should I use to present findings?
2             When and why the first method is used?
My intention is to apply a descriptive analysis by presenting: Frequencies, Mean and Standard Deviation of the questions them the total mean of each theme.
I really appreciate your help in this manner.
Best regards,
Amal
Relevant answer
Answer
Amal:
It isn't entirely clear to me what you are trying to do, but if you are asking about combining items, then what most people would do is sum the items, divide by the number of items, and then round to the nearest whole number. So if you take the mean of 5 items, 1.2 would be a 1, and 2.8 would be a 3. You then could report the percent of the sample that had a mean of 1, 2, etc. Sometimes people combine categories, such as 1 and 2 to mean disagree, and 4 and 5 to mean agree. It depends on what you are trying to show.
  • asked a question related to Survey Analysis
Question
4 answers
I'm looking at the National Diet and Nutrition Survey data. I'm interested in studying the association between some variables. My question is, although the survey is a 2 stage cluster survey, can i use the unweighted data as i'm not interested in the population parameters as much as i'm in the relation between variables. Must i use complex survey analysis or can i use normal analysis procedure. Again, I'm just interested in the relation bet some variables. The data set has several weighting variables to make results representative of the whole UK population.
Relevant answer
Answer
Dear Ahmed M. Kamel,
I Agree with Ahmed answer.
Thank you
  • asked a question related to Survey Analysis
Question
4 answers
I am trying to evaluate impact of an intervention that was implemented in very poor areas (more poor people, undeserved communities). In addition, the location of these areas were such that health services were limited because of various administrative reasons. Thus, the intervention areas had two problems: (1) individuals residing in these areas were mostly poor, illiterate and belonged to undeserved communities; (2) the geographical location of the area was also contributing to their vulnerability (as people with similar profile but living elsewhere (non-intervention areas) had better access to services. I have a cross sectional data about health service utilization from both types of areas at endline. There is no baseline data available for intervention and control. I am willing to do two analyses: (1) intent to treat analysis: Here, I wish to compare the service utilization in "areas" (irrespective of whether the household in intervention area was exposed to the intervention). The aim is to see whether the intervention could bring some change at "area" (village) level. My question is: can I use Propensity Score Analysis for this? (by matching intervention "areas" with control "areas" on aggregated values of covariates obtained from survey and Census?). For example, matching intervention areas with non-intervention areas in terms of % of poor households, % of illiterate population, etc. The second analysis is to examine the treatment effect: Here I am using Propensity score analysis at individual level (comparing those who were exposed in intervention areas with matched unexposed people from non-intervention areas). Is it right way of analysing data for my objective?
Relevant answer
Answer
Thanks a lot, Sebastian. This is very useful. I am working on it and I will update you once I address all these concern, to the extent possible with existing data available.
  • asked a question related to Survey Analysis
Question
8 answers
I am conducting a project in three different centers. I have sent questionnaire to 1,000 respondents in each center and got responses of as follows;
1) less than 200 in first two centers, and 
2) more than 750 in third center
As response of last center is extremely large as compare to first two. 
How I will compare these three centers data. Kindly guide me I want to apply t test and ANOVA.
thank you so much
Relevant answer
Answer
I think you need to do ANOVA  and to follow it with post hoc to compare between 3 groups. {if you hanlde your data with multiple comparison through ttest, then inflation in alpha will arise}
However, because you have one group obviosly larger in sample size, you need to consider: Weighted Means Analysis, in ANOVA, the attached word file is to help you in how to do weighted means, which was retrived from the following link. Hope this help
  • asked a question related to Survey Analysis
Question
14 answers
I would like to study the opinion of European citizens' on European integration, using a European survey with direct questions in merit and others such as income, preferred political party, and so forth. Then I would like to understand if there is a correlation between the results of this regression and the level of inequality within each country.
I am afraid of using the Gini index (or any other index based on income) to proxy inequality in this second regression, since there could be a problem of correlated variables. What could I possibly do? Is there an alternative index that I could use and that would still make sense?
Relevant answer
Answer
Maddalena,
Further, in case the Assets approach is of assistance, the following [my bold etc] has reached me, from R-G member  Alexander Dill
"World Social Capital Monitor
About
Revaluation of around 173 countries, their regions, their culture, their assets, their potentials - that's my task since 2007 when I published my "Global Freeware Index" for the first time.
Today I evaluate the major global indices, their unilaterism and their bias. And I'm testing new evaluation methods without aggregated data and binary questions.
I hope that the new method will not only being accepted and used but as well giving a new chance to countries in crises and conflicts."
Hoping this helps - Paul
  • asked a question related to Survey Analysis
Question
10 answers
I am doing research on FDI in retail.
To examine "consumer’s  perception towards FDI in retail." I had collected data from two cities meerut and agra. What type of test I should apply to analysis  the data. Hereby Questionnaire is attached. 
Relevant answer
Answer
Dear Sourabh,
I suggest you carry out a scale reliability analysis on each of your sections with at least 3 questions to see whether any of them could be reduced into a single scale variable. I recommend using a Principal Component Analysis to do this - see my guide:
Once you have done this you can either use a t-test or, if its assumptions are not met, a Mann-Whitney U test.
  • asked a question related to Survey Analysis
Question
5 answers
Dear all, in order to seperate the predictor from the criterion, I plan to collect data from two different groups of raters from the same company. Group A, 600 employees, will fill questionnaire 1 and Group B, also 600 employees, will fill questionnaire 2. 
1. How do I select these groups with minimum differences which might impact the cause-effect relationship? 
2. How do I measure or control for these differences. 
Thank you very much in advance 
Relevant answer
Answer
It sounds like you have two different experiments. The issue is not how Group A is different from Group B, rather it is how well Group A and Group B reflect the total population (all employees of that company). In that regards you might consider randomly selecting the first 600 from the total population, and giving them the questionnaire. Then select the second group of 600, but keep in mind that some of the people selected might have also taken the first questionnaire. Another approach would be to withdraw a sample of 1200 employees, and then randomly assign each employee to one of the questionnaires. You could also alternate: select the first employee at random for questionnaire 1, then select the second employee at random for questionnaire 2. Another option would be to give some people questionnaire 1, others questionnaire 2, and still others get both. Half of the ones that get both get questionnaire 1 first, the other half get questionnaire 2 first. What you do not want to do is to select the first 600 for questionnaire 1, and the second 600 for questionnaire 2. The reason is that your population for questionnaire 2 will be "all employees less the 600 removed for questionnaire 1." As each person is selected, you are changing the population of the company for the subsequent selection. Because a company is not a homogeneous entity, there is a risk that the process of selecting individuals will bias the outcome of the survey.
  • asked a question related to Survey Analysis
Question
5 answers
Likert scale used to asses attitude or opinion of participants. Reliability of such a scale is assessed by Cronbach's alpha test.
Now to create a knowledge score on vaccine preventable diseases. Each participants will be asked to identify the vaccine preventable diseases from 8 selected diseases. The response would be yes or no (two way close ended) . One score for correct answer and zero for wrong answer. Therefore there will be scale measuring knowledge in a range 0 to 8. The higher the number the higher the knowledge.
Any way to test reliabilty of such knowledge scales like Likert scale?
Relevant answer
Answer
Internal consistency with dichotomous items is a tricky thing.  I have attached a couple of articles I use when teaching psychometrics that are useful in laying out the problems (Sitjtsma) and some alternatives (Trizano-Hermosilla & Alvarado).
  • asked a question related to Survey Analysis
Question
2 answers
I have generally seen the scale used as a continuous scale. One recommendation for cut points (high hope, etc.) would be to use one or two standard deviations above or below mean, although this would be dependent on the characteristics of those who take the scale.
Substantively I am attempting to show change in the hope of individuals with serous mental illness at the time of onset of involvement in mental health services over time. I can use simple change scores but in presentation it would be clearer to be able to say that the individuals moved from low hope to ...
Many thanks
Relevant answer
Answer
I would stick with the continuous scale for analysis and modeling because of power issues and then turn the modeled results into caseness at the end for 'clinical' interpretation which you may get through conventions in the literature. And you should know the context of application better that anyone else - trust yourself on this.
  • asked a question related to Survey Analysis
Question
4 answers
I have conducted two identical surveys, 4 weeks apart, containing Likert scales. Some questions use the scale of "strongly agree" through to "strongly disagree" whilst the others use "very useful" to "not at all useful". The responses have been assigned numerical values in Excel and SPSS (e.g. strongly agree = 5). Both surveys were given to the same cohort of participants (as mentioned, 4 weeks apart).
The results were anonymous and no ID number was assigned to respondents. Hence, when I come to compare the results from the surveys (i.e. how the responses to the questions changed), I will be unable to match each individual's results. I was originally going to do the Wilcoxen signed rank test for my statistical analysis. However, I understand that this requires matched results (i.e. the responses from each individual directly aligned/in the same row as one another) which is not possible in this situation.
Has anyone got any advice? The only option at the moment appears to be the Mann whitney u test.
Thank you in advance
Relevant answer
Answer
What is the difference between having hypotheses and having research questions? The alternate hypothesis is typically "the research question".
  • asked a question related to Survey Analysis
Question
6 answers
Hello
I am a psychology student exploring the impact of a mindfulness program on stress, mindfulness practice, resilience and self-compassion. I have used the FFMQ, alongside other measurements. I am looking for a paper that interprets the scores and gives me suggested cut-offs.
I am aware that the FFMQ cannot be used as a whole as Observe does not correlate with the other domains.
Thank you in advance.
Nicola
Relevant answer
Answer
 I would love to get an answer to this question. Or at least studies where scores are « normalised » to get a sense of where the results lie... 
  • asked a question related to Survey Analysis
Question
2 answers
I'd like to do a rating scale survey to find out whether or not people in an (Voluntary) organization accept the authority from above. I am looking for items I can measure this with.  
Relevant answer
Answer
guess it depends on acceptance of authority in general or acceptance/liking of specific boss you have, but you might find something in Lincoln & Kallebergs 1990 'culture, control and commitment, cambridge UP
  • asked a question related to Survey Analysis
Question
7 answers
We have created a questionnaire  to measure prejudice and it is answer on a  5 points scale ( 1, agree to 5,disagree) , and I am trying to compare it to another questionnaire that is established and measures negative attitude. How do I run an analysis to find concurrent validity on SPSS? 
Relevant answer
Answer
To do concurrent validity, you may use 2 types of scales, one which convery the similar meaning to yours, thus you do convergent validity by doing correlation between the total scores for the 2 scales. If the correlation is high,,,almost .80 and above, then its validity is accepted.
On the other hand, you can do discriminant validity, by correlating your scale with another one with opposit themes. The result should show low correlation to say it is valid.
  • asked a question related to Survey Analysis
Question
1 answer
Hello,
I am doing my graduate research on Mincer model. How can I find the y0: earnings with no education and no experience. The data I will use will be my own distributed questionnaire data as micro data are not available for my country so can you please suggest what kind of data I should ask people for?
Relevant answer
Answer
Hi Zainab. There are two classes of variable you need to collect. Firstly, measures of formal education (years of schooling, qualifications etc). Second measures of experience such as age, working years etc). And of course wages and if possible hours of work, occupation and industry sector. Then you have the basic Mincer type model components. Wage = f (formal + informal human capital). Good luck. Cheers, Marc
  • asked a question related to Survey Analysis
Question
7 answers
I have found studies which use a qualitative approach, specifically using semi-structured interviews.  I am looking to find the problems which are faced by refugee entrepreneurs in Turkey based on which I would then propose some solutions.  Is it possible to do this through surveys? I have worked on a survey based on a previous study which performed qualitative research.  They created a model using different aspects which I then used to create a survey.  I am also in the process of doing semi-structured interviews.  I want to know how reliable would the survey results be in generalising the obstacles and proposing a solution? Should they be used separately or together in one paper?
Relevant answer
Answer
In my opinion, first start with focus groups (consists of relevant fields) and in-depth interview (as few as 5-6). The result will help you to construct a survey questionnaire which will be a good methodology for this research. The previous published papers, as you said, may help you to adopt or adapt few important items under respective dimension. Hope this will help you. 
  • asked a question related to Survey Analysis
Question
2 answers
I would like to use the Palmore's Facts on Aging Quiz for MSN students.  Does anyone know where to get permission to use this tool?
Relevant answer
Answer
Alex,
Thank you for your response.  Did you do a reliability test on this quiz?  I am looking for a Palmore quiz that has a Cronbach alpha coefficient.
Francine
  • asked a question related to Survey Analysis
Question
8 answers
My survey is on '' Lack of Critique and toughness of Sri Lankan English Literature-case at Advanced Technological Institute, Dehiwala, Sri Lanka''.The target population is the students of Higher National Diploma in English which is a two and half year diploma program. Those students are A/L qualified students.If any one is having some idea related to your local English Literature,kindly share with me to have some idea. 
Relevant answer
Answer
there are lots of qualitative methods, probably there is no "best" method but depend on the availability and so on. a thorough explaination of qualitative methods could be found on SAGE HAND BOOK - Qualitative methods, there are lots of varieties of Qual and you may read through and talk to people who are experienced in your research area about the practicality and the availability about them, talk to your supervisors are also a good way to find out more. 
All the best!
  • asked a question related to Survey Analysis
Question
2 answers
It seems Cohen's kappa requires that our unit (the sentence) be independent; however, the coded sentences are part of an interview, and so sentences follow each other along similar lines and the odds that a code will be repeated across several sentences is higher than by chance.
Also, our codes are not mutually exclusive, another condition of the kappa.
Which statistic should be used here instead?
Relevant answer
Answer
The key assumption in most measures of inter-rater reliability is that the ratings be independent. The units in the material being rated will almost never be truly independent, and that will apply to any statistic that you calculate.
In terms of choosing a coefficient to use, you could look into Krippendorff's alpha.
  • asked a question related to Survey Analysis
Question
40 answers
I am looking for an article or a book which support about 15 subjects is appropriate for testing reliability of a questionnaire (4 point Likert Scale with 26 items).
Thank you very much in advance for your kind advice. 
Relevant answer
Answer
Dear Suhathai
I don't think you could proceed a quantitative research with only 15 respondents. It's an insufficient amount of respondents even for a questionnaire pretest.
There's a relation between sample size and the number of items. Hair et al (1998) recommend between 5 and 10 respondents per item. In other hand Kline (2011) recommends at least 200 respondents when you work with structural equation modelling.
You should proceed the KMO test too. It will show you if the sample size is adequate. According to Malhotra (2006), anything above .6 show you that the sample size is adequate .
I guess you study nursing, is it correct? Did you ever thought about use Grounded Theory in your research?
I hope it can help you.
Best Regards
Yves
  • asked a question related to Survey Analysis
Question
5 answers
I am doing an analysis of a survey and people are part of either Group 1 or Group 2.  I would like to compare how these two groups respond to the main measure in my survey.  I was wondering what options I have for a quantitative analysis of this difference.
Relevant answer
Answer
Just to expand on David's answer, depending on the model you can also using other coding schemes: e.g., (-0.5, +0,5). There are are other reparameterizations that can aid interpretation for some purposes such as fitting a cell means style model.
  • asked a question related to Survey Analysis
Question
4 answers
Folks, I have used Qualtrics panels - not sure if you are familiar with this, Qualtrics have panels and charge a certain amount per survey depending on how scarce the panel respondents are - for example CEOs may be $50 a survey, whereas a middle manager might be $8-10 depending on criterion. It is a little difficult to get response rate data in this case. 
Another method some of my colleagues are using is  Amazon Mechanical Turk.  I have not tried this and am not familiar with the quality. It is definitely a less expensive option - I have heard it could be $1 per survey response. 
What are your views and experience on some good ways to get  quality survey data?
Relevant answer
Answer
Definitely i do recommend Qualtrics over MTurk for academic research;
Qualtrics will provide a dedicated project manager to help conduct, manage the survey access, distribute questionnaires and collect data from diverse demographic of the targeted sample, which will increase the quality of the collected data.
In contrast, MTurk is an academic research method being done online. However; this method may be leading to flawed or unreliable data.
 Dr. Kelly check the following video:
  • asked a question related to Survey Analysis
Question
2 answers
Hi, can anyone please inform me about a good survey for measuring the experience of visual work productivity? In particular I am interested in “visual work productivity”, such as reading and writing, in a visual ergonomic context. Kindly let me know of any related papers on this topic?
Relevant answer
Answer
dear sir
It has direct relation with productivity.  visual work & productivity has been studied during 90 & 2010 during this period font size of 12 is accepted as best font to be used for computer work. all such work established its importance.kindly consult ergonomics during that period you will get many studies. hope this will help you to understand the problems. pcg,chennai
  • asked a question related to Survey Analysis
Question
7 answers
I have identified a lot of different independent variables, too many to put all of them in one survey. I would like to split the survey in two sections,  part 1 is always the same while part 2 contains X random variables from a pool of all the other independent variables (that are not included in part 1). Can I still analyse this data in one model (e.g. a regression analysis)? My goal is to find out which of the independent variables explain most of the variance of my dependent variable. I would really aprreciate if someone can give me some ideas for methods or approaches that I can check out. Thank you!
Relevant answer
Answer
Dear David,
I am not an expert on multiple regression, but I hope that what I can offer is helpful. Forgive me if it's a bit colloquial - but maybe my being colloquial is of benefit.
First, somewhere in my stats background I have a notion that having a lot of independent (predictor) variables in a multiple regression is not only problematic statistically (I think it reduces your ability to get at what's going on because of problems with degrees of freedom) but also suggests that the researcher has not done enough prior thinking about what should be investigated / tested. In other words, it's not a good idea to throw a lot of things into the pot and stand back to see what bubbles up.
Furthermore, if you have a lot of independent variables you are probably likely to run into the problem of multicollinearity - i.e., that some of those independent variables are highly correlation with each other and should therefore not be included in the regression analysis because they will mess the multiple regression up.
I Googled these issues, and found the following site, which seems to include the concerns I have raised above:
I wonder, therefore, if you might be able to do some pilot testing to see if you can not only reduce the number of your independent variables but also, in the process, end up with a nice set of variables that you could ask everyone. This would involve identifying which of your variables are essentially getting at the same thing (i.e., multicollinearity), and then removing those that, in retrospect, you think are least good for your purposes.
One problem with what you are proposing is that you might end up with somewhat different samples for each set of multiple regressions and therefore your results might be dependent not only on the set of variables that you entered into the regression analysis but also the nature of the sample in each case.
I'm not 100% sure about what might happen with the missing data if you somehow combined your data, but my sixth sense is that a lot of missing data (it seems you'd have a lot) would create pretty enormous problems even if your stats package didn't buck and kick in their presence.
All the best with what you're doing.
Robert.
  • asked a question related to Survey Analysis
Question
4 answers
We are developing and validating a questionnaire regarding chronic disease and disaster preparedness. The questions are all categorical in nature ie: 
1) Which of the following items do you have as part of your household’s disaster preparedness? 
Check all that apply
 Water, two liters of water per person per day for at least three days
 Food, at minimum a three-day supply of non-perishable food
 First-aid kit
 Working flashlight
2) Do you require supplemental oxygen?
Mark only one box
If no go to question 5
 Yes
 No
3) Do you have 72 hours of supplemental oxygen cylinders (compressed gas, not liquid)?
Mark only one box
 Yes
 No
I am struggling to find the best way to examine test-retest. I think I should be using Kappa Stats but I'm a bit confused on it and there is little literature on the subject matter. 
Relevant answer
Answer
Hi Maddy
Steven is right in that the correct method of assessing test-retest reliability for categorical variables is kappa for Yes/No and nominal variables, and weighted kappa for ordinal variables such as Likert scales. You would like kappa to be at least 0.6 for each question.
Adrian 
  • asked a question related to Survey Analysis
Question
9 answers
I'm looking for something like the O'Brien article in Academic Medicine 2014 (Standards for Reporting Qualitative Research:A Synthesis of Recommendations) or CONSORT or something similar?
Relevant answer
Answer
Try looking at a survey research text.  Many will have a chapter on reporting results.  You might also look at the author guidelines for Public Opinion Research which is the journal for the American Association for Public Opinion Research (AAPOR) or the AAPOR website.
  • asked a question related to Survey Analysis
Question
7 answers
How many days between assessments is best for measuring test-retest reliability?
This is for a comprehensive structured interview for children and teens, to test the test-retest reliability (similar to the k-sads or scid for adults).
Relevant answer
Answer
Hello Amber,
I think you have received some excellent advice above along the lines of something that I wrote in a journal article about 6 years ago concerning test reliability.  Here's what I wrote then:
The difference in time should be large enough that respondents are not likely to remember or be influenced by their first set of responses when providing their second set but small enough that genuine differences in scores are not likely to have occurred. Any test-retest interval chosen should be justifiable relative to the stability that is believed to pertain to the construct of interest. ... Demonstrating temporal consistency through test-retest reliability is essential for three reasons. First, it protects researchers from assuming that their constructs are stable when they are inherently changeable. (Needless to say, it would not be appropriate to attempt to demonstrate test-retest reliability for a construct that is known to be labile.) Second, it ensures that instruments are not measuring things that are so unimportant or vague that people respond to them haphazardly. Third, it provides an opportunity to explore reactivity, which refers to the situation in which people’s exposure to an instrument at one point in time influences the way they respond on a second occasion.
One additional thing that you might consider is how you are going to assess your test-retest reliability, e.g., whether you should use a simple correlation or rather something like an intraclass correlation coefficient (ICC), or something else. To some extent that will depend on the nature of the data that you collect. If you do use an ICC, I think it should be ICC(3,1) absolute agreement. Please let me know if you'd like more information about this. ICCs can be a bit tricky, and I've just written a journal article about them.
Every good wish with your research.
Robert.
  • asked a question related to Survey Analysis
Question
5 answers
I am starting my dissertation related to spiritual well being and compassion fatigue, and I plan to use three different measurement surveys, such as the Professional Quality of Life (ProQOL) measure. What is the best way or ways to make comparisons, statistically among the three scales?
Relevant answer
Answer
Hello Alan,
I am responding primarily to your most recent question:
It all depends on what you mean by "mixed methods".  By these words, many people assume that the distinction between methods refers to what are traditionally referred to as qualitative methods versus quantitative methods.
Given that your original question indicated that you were interested in analysing your data statistically, that essentially rules out qualitative methods.
In essence, if you use surveys / instruments that yield only quantitative data (numbers), as does the ProQOL, you have ruled out using mixed methods in the traditional sense. (i.e., you have ruled out qualitative methods.) However, it depends on what you mean by "methods".  There are different methods within the quantitative bundle.  Up above, Peter has referred to correlations.  They are related to other things such as regression analyses.  But there are other quantitative "methods" such as tests that assess differences. These include t-tests and ANOVAs. 
Much depends on what your surveys are like in terms of the methods you can use with them. Perhaps some of your surveys yield data (usually verbal responses) that permit qualitative analyses. However, even verbal data can be analysed in ways that are primarily quantitative - and that therefore permit subsequent statistical analyses.
Apart from that, ultimately, a lot depends on what your research questions are.  For example, are you wanting to make comparisons between the scales, or are you wanting to make comparisons between groups within your research based on their results on each of the scales?  These are slightly different focuses.
Robert.
  • asked a question related to Survey Analysis
Question
7 answers
Dear researchers and Professors
I am going to run exploratory factor analysis for a questionnaire with 16 items. Prior to running an EFA, it was calculated the internal consistency with Cronbach's alpha. Alpha was estimated .41!
I have two questions:
1-     Can I run that EFA?
2-     If your answer is “No” (because of low alpha and inadequate internal consistency), what is your comment?
Sincerely yours
Hamid  
Relevant answer
Answer
You can certainly run the EFA. One possible reason for a low alpha is that your measures do not all fall on the same scale, which would show up in the EFA as two or more distinct factors.
Before doing that, however, I would check for two other common problems (which can both affect your EFA).
1. You have negatively correlated items that need to be reversed.
2. You have problems with missing data so that some of your correlations are very low.
  • asked a question related to Survey Analysis
Question
4 answers
I have 4 items to a questionnaire in which the answers were the same type of scale: from 0 to 5, 0 for always false and 5 for always true. I want to study if an item score can determine certain scores on other items.
Relevant answer
Answer
What do you mean the data is qualitative? You have 4 items that have responses from "Always false" to "Always true"  ranging from 0 to 5. That sounds quantitative. 
Cronbach's Alpha would give you a summary of the internal consistency of the items as if they were part of a single scale. With such a small number of items and participants, however, I would be worried about huge fluctuations. You could try it. You could also look at the mean and variance of the intercorrelations. 
  • asked a question related to Survey Analysis
Question
4 answers
Variable in my dataset are item responses to multiple scales.
Relevant answer
Answer
Agree. Variables with missing values should usually be imputed jointly (i.e., all at once) in order to preserve the associations between them. Of course, you do not necessarily need to impute the whole of your 750 variables. Just include those variables in your multiple imputation model which are also involved in your analysis model, i.e. which are thought to affect you dependent variable. Some practical advice on how to select variables for your imputation model can be found in the attached paper. 
Regards
  • asked a question related to Survey Analysis
Question
2 answers
I want to conduct quantitative analysis of MMPs and TNF-alpha. Can anyone suggest better and reasonably cheaper method or assay?
MMP-8,9,13 anf TNF-alpha
Relevant answer
Answer
Thanks for suggestion
  • asked a question related to Survey Analysis
Question
3 answers
Hi Colleagues,
Please advise on models or procedures (literature) I can consult for using a matching procedure to assign participants to a treatment and control group in a quasi-experiment. I'm aware that due to non-randomization, selection bias is a major threat to the internal validity of my research. Therefore, I wish to control for any biases of covariates in the design and data analysis. I've come across pair matching, stratification, and covariate adjustment but have found no practical procedure as to how I can propose to implement these models in my study. I know that matching participants to a control and treatment group, one must ensure that participants have a great deal of 'similarity' to reduce the non-equivalence.
So, simply put: Do you know of any practical models that i can use in my research to do matching or participants in a quasi-experiment?
Relevant answer
Answer
Check prospensity score matching ( modeling the group assingnment using pre interventional covariates).
This is a great introduction:
Hope that helps, best wishes, Manuel
  • asked a question related to Survey Analysis
Question
3 answers
I want to create an survey in in which I will put a person's picture and demand the respondent to guess his/her age. Then I will introduce an offer ("the average of guesses about his/her age is.....") and let him change his answer after this offer. So, how to create such a survey in which the respondent can see his previous answer and can change it. Thanks in advance.
Relevant answer
Answer
Create two sheets in an online form (e.g. google dicuments) differing in the presence of the "offer" only.  As an another suggestion, you may also check wix.com (a very intuitive and easy web builder) where you can create a customized online form with great degree of freedom :)
Best regards, M.
  • asked a question related to Survey Analysis
Question
12 answers
Currently analyzing the data from my questionnaire. There I do the first step of a scale development process. I have around 45 items which I submit to exploratory factor analysis. Respondents answered all items for four countries. My question: When I run the EFA, should I run it separately for the answers of all four countries?
When I consider all answers of all items for the EFA I would have a within subject design as one person answered each item for all four countries.
So should I run four separate EFA's for each country's answers and then compare the results? What should I do if I receive different factor structures?
Relevant answer
Answer
I would encourage you to use multi-group exploratory structural equation modeling (ESEM, it ist not the same as EFA!!!). As Marsh et al. (2011) state, ESEM is an integrative framework of CFA, SEM and EFA. It is a comprehensive approach which would allow you to model the data of your countries simultanously, without running seperate EFA with the direct possibility to test invariance and gain standard indicators of model fit. ESEM integrates interesting features like the possibility to add parameter constraints and testing invariance (e.g. constraining loadings equal across countries to establish week invariance, and add equality constrains on the intercepts to establish strong invariance). In addition, it is more flexible than CFA since the independent cluster assumption inherent in CFA is often very restrictive, especially if you want to do something within an exploratory framework. Since the number of clusters is low (only 4), multilevel EFA is probably not appropriate (Linda Muthen recommended a minimum of 30 - 50 clusters for ML-CFA: http://www.statmodel2.com/discussion/messages/12/86.html?1380849285). 
Check the paper of Marsh, Asparouhuv and others to gain insight in ESEM:
Marsh, H. W., Nagengast, B., & Morin, A. J. (2013). Measurement invariance of big-five factors over the life span: ESEM tests of gender, age, plasticity, maturity, and la dolce vita effects. Developmental psychology, 49(6), 1194.
Asparouhov, T., & Muthén, B. (2009). Exploratory structural equation modeling. Structural Equation Modeling: A Multidisciplinary Journal, 16(3), 397-438.
Marsh, H. W., Morin, A. J., Parker, P. D., & Kaur, G. (2014). Exploratory structural equation modeling: An integration of the best features of exploratory and confirmatory factor analysis. Annual review of clinical psychology, 10, 85-110.
Marsh, H. W., Liem, G. A. D., Martin, A. J., Morin, A. J., & Nagengast, B. (2011). Methodological measurement fruitfulness of exploratory structural equation modeling (ESEM): New approaches to key substantive issues in motivation and engagement. Journal of Psychoeducational Assessment, 29(4), 322-346.
Here a paper that used ESEM for cross country comparisons:
Tomás, I., Marsh, H. W., González-Romá, V., Valls, V., & Nagengast, B. (2014). Testing measurement invariance across Spanish and English versions of the physical self-description questionnaire: An application of exploratory structural equation modeling. Journal of sport & exercise psychology, 36(2), 179-188.
Hope that helps, Manuel
  • asked a question related to Survey Analysis
Question
7 answers
Specifically, ½ the sample was a stratified random sample of peri-urban households and ½ was a hand-selected sample of surrounding villages where all households were surveyed. I would like to do one regression analysis with this data so that I have enough power for significant results. However, I am concerned that the difference in sampling methods would invalidate the analysis or undermine its interpretation. Can anyone advise? Is it valid for me to put these two groups with different sampling methods into one analysis? How might that affect the interpretation of the analysis?
Relevant answer
Answer
Thank you Jim and Vincent. Your answers have been very helpful.
Jim- I will look into the reference you suggested. I should be able to pull it from my library here.
Vincent - Unfortunately, the two areas sampled are statistically different on a number of variables, including the outcome variable.
I'll push on. If I find a different solution, I will follow-up this post.
Kind regards.
  • asked a question related to Survey Analysis
Question
4 answers
I am working on my thesis and I have a survey that has true/false and multiple choice questions. From what I understand, KR20 is used for dichotomous variables to estimate reliability coefficient. Also, if multiple choice questions are not equally weighted, I can use KR20.  Am I correct in doing KR20 for both trues/false and multiple choice items? Any suggestions and input is greatly appreciated.  If you can suggest references that would be great.  Thank you!
Relevant answer
Answer
Thank you very much for all your responses!!!!!
With regards to the multiple items, I re-coded them into 0 and 1 as correct or incorrect, and now they are dichotomous as well.
Dr. McLean.
Thank you for your response.  You suggested factor analysis.  Can I do factor analysis of dichotomous variables? If so, what procedures should I use?
Thank you so much for your time.
Tatyana Aultman
  • asked a question related to Survey Analysis
Question
10 answers
Hi, I have run a questionnaire with 9 dimensions, each of them. Because they are measured using a 7-scale Likert scale, they are ordinal variables (please correct me if I am wrong). My hypothesis claims that there is a relationship between the dimensions, so I need to group the items into dimensions to be able to correlate them. I seem to remember that can be done with the mean with scale variables but not with the ordinal ones. 
How can I group them then to check if there is a relationship between variables?
Thanks in advance!
Relevant answer
Answer
I have to say that I am somewhat confused by all the recommendations to use various forms of rank-ordered correlations. The use of Pearson's correlation is well-established with Likert-scored items, via Cronbach's alpha. I suspect that you are getting a lot of recommendations based on "theory" rather than on typical practice.
So, I suggest that you check what the standard practices are in your own field, before making a decisions either way.
  • asked a question related to Survey Analysis
Question
1 answer
Hello everyone,
I have previously used a shortened IPIP scale that consists of 30 items - I have lost my copy and I am having real trouble finding it...Does anyone have a reference or link to how it is scored?
Thanks 
Relevant answer
Answer
Hello,
The IPIP is a pool of 3,000+ personality items that can be combined into many different scales intended to measure different constructs. You're probably looking for the short form of a scale for one of those constructs