Week 7: Statistical Analysis

Page 1

Descriptive analysis technique:

- ensures data is correct

-validity and reliability

Validity: are we measuring what we are trying to measure?

data collection - was the procedures followed or biases present?

external consistency - transferable results

Internal consistency - Does x cause y?

Statistical conclusion credibility - is the interpretation appropriate?

Reliability: are your results accurate?

-negatively phrased questions throughout a survey can help determine this

Look for:

are the results free from errors?

was an appropriate collection measure used?

appropriate and accurate?

does the sample represent the population?

Data validation:

fraud

screening: screen out certain incomes. Do a validation call back

procedure: callback to ensure correct method was used.

Completeness: ensures all questions were asked

courtesy: checks the pleasantness of the interview

Data Editing:

essentially checking for mistakes. PART OF SPSF

-correct questions asked

-proper answer recording

-correctly screened respondent

-proper recording and coding

Recording and coding:

assigning values to responses.

Single response: pre-determined category eg. 18-30 yr old

open response: pre-determined category

Recall response: count data

Missing data solutions:

1. replace missing data with the mean of the other data

2. delete data set.

Preliminary data analysis

data description

tabulation

Frequency distribution

Central tendency etc.

Variance:

the mean for all observations/ (number of observations -1)

Standard deviation: square root of the variance

Coefficient of variation: SD/mean

Sd=1 includes 65% of sample

=2 includes 95%

=3 includes 99%

If outside of this you either have a non representative sample of a sampling error.

Distribution of shape:

skewness - is it equal around the mean?

Ktosis - how flat/tall is the peak?

>0 leptokurtic ie. tall

<0 platykurtice. flat.