Unit 9 introduction to ih stat
- 1. Introduction to IHStat software(requiredfor assignment 3)
- 2. IHStat is a free statistical package developed by the American Industrial Hygiene Association for evaluating occupational samplesThe program is a preprogramed Excel worksheet that:computes descriptive statisticsdetermines if the distributions are normal or log normally distributedcalculates associated confidence limitsIdeal for determining compliance with little workWhat is ihstat?
- 3. Normal distribution When data resembles a bell shape (i.e. equally clustered around the mean/average)Often occurs with noise data and/or physiological monitoring (ex. blood pressure, heart rate)Log Normal distribution When original data is skewed and does not resemble a bell curve Occurs frequently with air monitored sampling (e.g. particulates)Transforming the data (i.e. taking the natural logarithm of each sample) provides us with a normally distributed variableLognormal vs normal distributions
- 4. The remainder of this PowerPoint will walk you through how to use IHStatwith the use of screen shots.Before we can proceed you need to:Download and open IHStat from Moodle or from the AIHA website Ensure macros are enabled in ExcelNOTE: if you are unable to enable macros there is a macro free version available on the AIHA website; however, the layout and design is slightly different.How do I use Ihstat?
- 5. Opening page of ihstat programSTEP 2:Click button to enter programSTEP 1: Select language preference
- 6. How to enter exposure limit for comparisonsSTEP 3:Enter /change occupational exposure limit of hazard you are evaluating by typing in the new value.For example: chlorine has a TLV of 0.5 ppm.
- 7. How to clear dataSTEP 4:When you open IHSTAT there may be existing data in the sheet.Clear existing data to input new data by clicking on the eraser (circled).
- 8. How to Enter dataSTEP 5:Enter your dataCopy the data points from your Excel sheet you want to evaluate Data must be copied from one columnGo to IHStat program, click the first empty cell under the “Sample data” column and paste data (or enter by hand)When you paste your data the statistics/ graphics will be computedNOTE: Max 200 data points can be entered into programdescriptive statisticsStep 6Recording descriptive statisticsThis first box provides the descriptive statistics you will want to include into your reports such as the number of samples, mean, standard deviation, and geometric mean and standard deviation (i.e. mean of log transformed data) , etc.
- 9. You will also note they include the percent of samples above the OEL you indicated in Step 3evaluating datadistributionsStep 7:Determine the distribution of your dataThe second data box helps you determine if your data fits a normal (i.e. bell curve) or a lognormal distribution (i.e. skewed curve).This is important as the confidence limits used to determine compliance are computed differently.IHStat makes determining distributions easy. They provide you with the W-test statistic, but also provide you with a simple /easy way to interpret the results:Yes/blue thumbs up = fits distribution
- 10. No/red thumbs down = does not fit distributionIn this example we can see that the data fits a log normal distribution.
- 11. α levelNOTE:α = 0.05 is referring to the confidence levelα = 0.05 means we have a confidence level of 95%IHStat uses 95% confidence level for all tests
- 12. Choosing confidence limitsWhat do you do in cases where:Data fits both a normal and lognormal distributionIn this situation you can choose either option however, a normal distribution is easier to interpretData does not fit a normal or lognormal distributionIn this type of situation you are best using a nonparametric analysis method, which cannot be done in IHStatobtaining confidence limitsStep 8:Obtaining confidence limits (i.e. UCL, LCL)Now that you know which distribution your data follows you can select the confidence limits provided in either the:Lognormal parametric statistics boxUse this box if your data fits a lognormal distribution from Step 7
- 13. Confidence limits are calculated by Land’s Exact methodNormal parametric statistics boxUse this box if your data fits a normal distribution from Step 7
- 14. Confidence limits are calculated by t-statistics12
- 15. saving filesYou can save the program with the current data by clicking the save icon on the program sheet.NOTE: this will save the file to a default location and does not allow you to change the name of the file so you continually overwrite the existing data.When you are doing your assignment it is a good idea to save each analysis (i.e. each grouping) as a separate file so you don’t have to reenter data if you want to double check your numbers. To save each analysis as it’s own file go into the file menu and select “save as”Imbedded notes/commentsPlease note that in some of the data result boxes there are small red triangles. These indicate that there are comments associated with the test
- 16. To view comments hover over triangle and the comment will appear (as shown above)
- 17. These comments are meant to provide you with information or descriptions of the statistic