I have data from two national groups on familiarity of travel destinations. Respondents indicated ‘familiar’ or ‘not familiar’ for each destination. Then for each destination answered 4 statements on a 7 point likert scale.
I want to compare familiarity Of the destinations overall and the difference in familiarity between the two national groups. As categorical data is Chi2 the best for this purpose?
As for the 7 point Likert statements, best to average the scores and independent t test or ANOVA？ or is there a better test.
thank you for your help and suggestions.
When I am read your question I see that you have response variable is ( o,1) in this case you can use categorical data and Loigistic regression ...
if you need any information you ask me..
Thank you
Regarding your first question, yes, a chi-square test could be used as it is optimal for categorical data. Be aware that will only provide you with the correlation between the dependent variable. If you would like to assess the strength of association then you could further go with logistic regression.
For the second question, the test to propose will depend on what you would like to know. If you can come clear on that. I could make some proposals. FYI, t test or ANOVA are used for numeric/scale variables and not categorical ones.... Hope this helps. Cheers!
Hi,
According to your first question, I think better to do regression analysis to compare familiarity of the destinations overall and use pearson correlation for to determine the difference in familiarity between the two national groups.
For the 7 point statements, use ANOVA test to determine differentiation.
@Emmanuel and @Fadhil thank you for your suggestion to use logistic regression. Do you know of any good examples of usage as a reference? The main things i want to explore are the difference in recognition between the different nationalities, previous visit experience to the destination (yes/no) and participants knowledge of the language of the country the destinations are located in. From photographic evidence participants identified the name of the destination and the city where is it located. I coded with '1' for recognition and '0 for 'no recogntion'. Hiroko Sata
Sorry had a technical problem. To continue... participants also indicated if they had any visitor experience in each destination or not (again coded 1 or 0). Participants also responded to four 7 point Likert for each photo. This is the data on familiarity. @Yohan is it ok to use pearson for this given it is discrete data and only 7 points? I had about 15% empty data cells on the Likerts - more than expected - is there a good way to include this data or best to ignore in the analysis? also i had a lot of mid point data (4) so had thought it might be best to ignore the mid point, and consolidate the rest of the data to 'positive' and 'negative' responses and use that instead as a category rather than the raw averages and resultant standard deviation, for use in anova etc. Once again, BIG thanks to everyone for your suggestions.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA