Mind map
SPSS 2: item analysis and cronbach's alpha
Map content
SPSS 2: item analysis and cronbach's alpha
yields a table of correlations between each item and the overall total for the questionnaire. (item total correlations)
- cronbacks alpha - table of values showing what each alpha would be if deleted.
negative item-total correlations = either recode to see if it makes an improvement; if still not then delete.
- small item-total correlations and poor reliablility = if alpha is low (below 0.7) you need to delete the items with the smallest contribution to see if it affects overall score
- if overall score is 0.7 keep items which are weak
- only re-code an item once
- small item-total correlations and poor reliablility = if alpha is low (below 0.7) you need to delete the items with the smallest contribution to see if it affects overall score
SPSS step one = recode intentionally negative items. Transform > recode into a different variable
- SPSS step two: initial Cronbach's alpha. Analyze > Scale> Reliability Analysis. Statistics > Scale if item deleted. continue
- SPSS step three: further recoding. WORK WITH ONE ITEM AT A TIME. 1st - recode the item with the largest negative item-correlation. repeat until 0.7.
- SPSS step four: DELETING ITEMS. start with highest alpha if deleted. continue until 0.7
- ties? if same cronbach's when deleted look at the question wording or lowest item-total correlation.
- SPSS step four: DELETING ITEMS. start with highest alpha if deleted. continue until 0.7
- SPSS step three: further recoding. WORK WITH ONE ITEM AT A TIME. 1st - recode the item with the largest negative item-correlation. repeat until 0.7.
- SPSS step two: initial Cronbach's alpha. Analyze > Scale> Reliability Analysis. Statistics > Scale if item deleted. continue
Correlation
- SPSS step one:create a scatter plot: graphs> chart builder.
- SPSS step two- choose the chart. e.g. scatter, select and drag to preview area.
- from 'Variables' panel drag and drop each variable onto the axis
- SPSS step two- choose the chart. e.g. scatter, select and drag to preview area.
- correlational analysis. Analyze> Correlate> Bivariate
- Bivariate correlations. select dependent variavles and move into variables panel
- choose which correlation test to run, Pearsons or Spearmans. choose two tailed or one tailed.
- writing up = r=.733, N=25, p<.001. The relationshup accounts for 54%
- PARTIAL CORRELATION: Analyze> Correlate> Partial
- 1st = dependent variavle into variables. then choose which variable to control.
- PARTIAL CORRELATION: Analyze> Correlate> Partial
- writing up = r=.733, N=25, p<.001. The relationshup accounts for 54%
- choose which correlation test to run, Pearsons or Spearmans. choose two tailed or one tailed.
- Bivariate correlations. select dependent variavles and move into variables panel
- multiple correlation = adding all the variables into the dialog box at once.
- SPSS step one:create a scatter plot: graphs> chart builder.
Images and files in the map
Information
- Description
- undergraduate Research Methods Mind Map on SPSS 2: item analysis and cronbach's alpha, created by Katie gavriel on 18/03/2019.
- Created by
- Katie gavriel
- Views
- 2
- Rating
More on Validity, Reliability and Bias