Labour Market in the Czech Republic

Use of annex tables

Tab I Estimates of 95% confidence interval of the basic estimates for population aged 15+ (thousand)

Variants: 1a for quarterly estimates in total
1b for quarterly estimates for men and women
2a for annual estimates in total
2b for annual estimates for men and women

The table is designed to establish an approximate 95% confidence interval of basic estimates for the basic sample of 15+ population in the whole country and all its regions. For instance, if we want to know the confidence of the estimated number of employed men in the 4th quarter 2001 (2703.2, thousand in the Czech Republic), we shall find in the table 1b a row next to the number 2703.2 in the column of the Czech Republic. This is 25.1 thousand for the estimate size 2600 thousand. The next neighbouring value – 24.5 thousand - corresponds to the estimate 2800 thousand. Since the difference between 2703.2 and 2600 makes up roughly one half of the difference between 2800 and 2600, we shall subtract one half of the difference between 25.1 and 24.5 from 25.1 and get 24.8 in the end. This means the 95% confidence interval for the estimate of the number of employed men in the 4th quarter 2001 is approx.. 2703.224.8 thousand, i.e. there is a 95% probability that the actual number of employed men in the Czech Republic was not lower than 2678.4 thousand and not higher than 2728.0 thousand.

Tab II Estimates of 95% confidence interval of the partial estimates for population aged 15+ at the national level

Variants: 1a for annual estimates in total
1b for annual estimates for men and women

The table is designed to determine an approximate 95% confidence interval of partial estimates for the basic sample of 15+ population at the level of the Czech Republic only. For instance, if we wish to establish the confidence of an estimate of the unemployed with the university education attained in 2001, who stood at 14.7 thousand out of the total of 421.0 thousand unemployed persons (3.5% of all of the unemployed), we use the Table to find the value in a row approximately corresponding to 421.0 and in a column approximately corresponding to 3.5. We can also make the following correction by a simple linear interpolation:

	3	3.5	4
400	0.53		0.60
421.0	cca 0,52 =0,53-(421-400)/(450-400)* (0,53-0,50)	cca 0,56 =0,52+ (3,5-3)/(4-3)* (0,59-0,52)	cca 0,59 =0,60-(421-400)/(450-400)* (0,60-0,57)
450	0.50		0.57

This implies that there is a 95% probability that there were not fewer than 3.5%-0.56% (12.4 thous.) and more than 3.5%+0.56% (17.1 thous.) of the unemployed with the university education attained.

Table II can also be used for basic aggregates in the age groups and sex for the whole country, provided the basic aggregate is replaced with the size of the basic sample and partial aggregate with the respective estimate.

In this chapter, we intended to give the reader general instructions on how to roughly determine the error which arises from applying characteristics of the sample to the basic sample. This error depends on three variables (on four in the case of partial aggregates), namely the size of the sample and of the estimate and, to a lesser extent, on the size of the basic sample. Giving an objective overview on errors of all the estimates would require compiling a large annex of tables and it would be difficult for common readers of economic publications to find the necessary information there. This is why all of the methods used are considerably approximate but still fully sufficient for getting an idea of the accuracy of the estimates.