lecture 2

lecture 2

1. parametric assumptions: How to test parametric assumptions using spss ---> Kolmogorov Smirnov test and Lvenes test

2. parametric assumptions

   • parametric tests are based on assumptions about the distribution of measures in the population. A normal distribution is usually assumed
   • If comparing more than one group the groups should have equal variance
   • It is recommended to use parametric tests whenever possible. Most are quite robust and limitations well documented
   • 1. The data flow of normal distribution Normal distribution - a particular pattern in a set of numbers where the avergae number is seen most often and the further a number is from the average, the less often it is seen. A true normal distribution follows a very particular bell shaped curve.
   • 2. The data are at the interval or ratio level of measurements: Numbers that are at the interval or ratio level, that is, proper numbers, like those you deal with everyday - have more info in them than other numbers. As these numbers have more info, its not surprising that more powerful parametric tests insists that you only use numbers of this type. You can;t do parametric tests if you use ordinal numbers such as finishing positions in a race.
   • 3. Homogenity of variance - both sets of numbers should be spread out in the same way. When you have more than one set of data e.g when you are comparing two groups of people - parametric tetsts assume that for each group, the scores are spread out in a similar way. This simply means that if you look at a measure of dispersion for the groups, such as standard deviation, this should be more or less the same

3. To test normality of data we can use a Kolmogorov -Smirnov test. This determines the likelihood of the data belonging to normal distribution

   • tests the null hypothesis that the distribution is normal. If signoificant then the data is NOT normally distributed

4. To test for homeogenity of variance we can use a levenes test

   • This determines the data sets are from the same population
   • A value greater than .05 means that the variability in your two conditions is about the same

5. Normal distribution do not necessarily have the same means and standard deviations. A normal distribution with a mean of 0 and a standard deviation of 1 is called a standard normal distribution. Areas of the normal distribution are often represented by tables of the standard normal distribution.