Dear Megan, You have mentioned that Your questionnaires have 5 questions for measuring awareness and 10 questions for measuring their attitudes towards dyslexia. You can give scoring according to level of awareness for each questions and generating total score for the awareness gives you one single variable for measuring awareness. Then either you can categorize this total score or use it as continuous variable as a predictor variable (IV). Similarly you can make a new variables explaining you attitude towards dyslexia and it will be your Response variable (DV). Then you can proceed for your analysis.

Agreeing with Ujjaval, I’d like to add that probably the best statistical approach here is to use regression analysis, predicting attitudes (DV) from awareness (IV). The advantage of regression is that you can easily adjust the effect of interest for possible confounding factors, e.g. by including age and gender as additional covariates. Also, I don’t know your items and/or the instruments they are probably coming from, but it seems to me that while awareness may pretty much be unidimensional (one “scale” ranging from low to high), attitudes may not necessarily have to be. Maybe you items form two or more unobservable facets of attitudes. Then you would not have one DV but two or more DVs representing different aspects of attitudes. You could test whether your items on attitudes form a unidimensional or a multidimensional construct by using factor analysis. However, this would require a rather large samples size (approx. 10 observations per items, i.e. in your case at least n=100) and may therefore not be applicable for your study.

There are good tutorials for both regression and factor analysis in SPSS as well:

https://www.spss-tutorials.com/regression/

https://stats.idre.ucla.edu/spss/output/factor-analysis/

Hope this helps.

Best regards

Dear Megan,

Following the suggestions from Ujjaval Srivastava and Alexander Pabst , it is more acceptable to use a dimension reduction technique to create your DV and IV. My advice would be to run a principal component analysis with your 5 awareness questions followed by the same analysis with your 10 attitudes questions. Save the regression scores for the first component with each analysis then use this in your correlation or regression model. You may need to re-run this analysis if some of the item loadings are too small. There is detailed advice on sample size and the size of the first eigenvalue in a paper by Halil Yurdugül (see my study guides).

When carrying out a correlation or regression, start with a scatter plot to assess whether the parametric assumptions are met.