Scribd
Upload
6 views3 pages

Normality Test for Daily Production Data

This document provides steps to conduct a Kolmogorov-Smirnov test to determine if a data set is normally distributed. It includes: 1) Stating the null and alternative hypotheses that daily production data is or is not normally distributed. 2) Identifying the decision rule to reject the null hypothesis if the calculated test statistic Dcalc is greater than the critical value Dcrit. 3) Calculating Dcalc by ordering the data, identifying identical values, constructing tables of the empirical and theoretical distributions, and taking the maximum distance between the two. 4) Making a decision not to reject the null hypothesis since Dcalc is less than Dcrit, indicating the data is likely normally distributed.

Uploaded by

PeterMajor
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views3 pages

Normality Test for Daily Production Data

This document provides steps to conduct a Kolmogorov-Smirnov test to determine if a data set is normally distributed. It includes: 1) Stating the null and alternative hypotheses that daily production data is or is not normally distributed. 2) Identifying the decision rule to reject the null hypothesis if the calculated test statistic Dcalc is greater than the critical value Dcrit. 3) Calculating Dcalc by ordering the data, identifying identical values, constructing tables of the empirical and theoretical distributions, and taking the maximum distance between the two. 4) Making a decision not to reject the null hypothesis since Dcalc is less than Dcrit, indicating the data is likely normally distributed.

Uploaded by

PeterMajor
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Tutorial 5: Question 4.

Note: ALWAYS FIND D VALUES FOR BOTH SMIRNOV and LILLIEFORS FROM LECTURE SLIDES!!! I accidently use the textbook tables for Q3, to get Dcalc = 0.410, it should be 0.409 from the lecture slides.
Step 1: Hypothesis: H0: Daily production is normally distributed H1: Daily production is not normally distributed Step 2: Decision Rule: Reject H0 if Dcalc > Dcrit = D,n =D0.05,15 = 0.220 Step 3: Calculated test statistic (Dcalc) Given the initial data: 93 101 103 88 95 98 101 94 91 101 105 92 96 95 94

Step i) Order data smallest to largest 88, 91, 92, 93, 94, 94, 95, 95, 96, 98, 101, 101, 101, 103, 105 Step ii) Are any values the same? YES!!! In both the Lilliefors and KolmogorovSmirnov it is critical to identify any identical values!!!! (In Q3, there were no identical values of IQ, so it was not an issue) however when we have multiple values we group them together into the one value and this effects the S(x) = j/n column. S(x) = j/n Where j is the number of values that are less than or equal to that value. Example: Q. In the data set above, how many values are less than or equal to 88? A. 1 only 88 is less than or equal to 88 therefore j = 1 Thus S(x) for x = 88 is S(88) = 1/15 = 0.0667 Q. In the data set above, how many values are less than or equal to 91? A. 2 88 and 91 are both less than or equal to 91 therefore j = 2 Thus S(x) for x = 91 is S(91) = 2/15 = 0.0033 Now lets try an example where they are two or more values that are identical. Q. In the data set above, how many values are less than or equal to 94? A. 6 88, 91, 92, 93, 94, 94 therefore j = 6 Thus S(x) for x = 94 is S(94) = 6/15 = 0.4

Step C: Construct the table: Remember we dont list multiple occurrences in the table, but we must remember how many there were.

Count (not necessary, but can be helpful) 88 91 92 93 94 95 96 98 101 103 105 (1) (1) (1) (1) (2) (2) (1) (1) (3) (1) (1)

-1.7446 -1.1264 -0.9204 -0.7143 -0.5083 -0.3022 -0.0962 0.3160 0.9341 1.3462 1.7583

0.0405 = 0.5 0.4591 0.1292 = 0.5 0.3708 0.1788 = 0.5 0.3212 0.2389 = 0.5 0.2611 0.3050 = 0.5 0.195

1/15 =0.0667 2/15 = 0.1333 3/15 = 0.2000 4/15 = 0.2667 6/15 = 0.4000 8/15 = 0.5333 9/15 = 0.6000 10/15 = 0.6667 13/15 = 0.8667 14/15 = 0.9333 15/15 = 1.0000

0.0257 0.0041 0.0212 0.0277 0.095 0.1521 0.1398 0.0412 0.0429 0.0218 0.0392 Dcalc = Max |F(x)-S(x)| = 0.1521

0.3821 = 0.5 - 0.1179 0.4602 = 0.5 - 0.0398 0.6255 = 0.5 + 0.1255 0.8238 = 0.5 + 0.3238 0.9115 = 0.5 + 0.4115 0.9608 = 0.5 + 0.4608
-

Note: A lot of people coming to consult have been using the rule 0.5 (probability from table) to find all their F(x) values! This is WRONG!!!!! We can only do that for a negative Z score! We are looking for the CUMULATIVE probability, so for positive values..

The associated area is thus, 0.5 + 0.4115 = 0.9109.

Z = 1.35

Step 4: Decision Decision: Do not reject H0 as Dcalc < Dcrit Step 5: Conclusion There is insufficient evidence at the 5% level of significance to conclude that daily production is not normally distributed.

You might also like