Random variables that are inherently positive occur in many real life applications. The suitability of the lognormal distribution has been validated for several such applicatio in particular, for analyzing biological data and for analyzing data on workplace exposure to contaminants. A problem of interest in this context is statistical inference concerning the mean of the lognormal distribution. For obtaining confidence intervals and tests for a single lognormal mean, the available small sample procedures are based on a certain conditional distribution, and are computationally very involved. Occupational hygienists have in fact pointed out the difficulties in applying these procedures. For two independent lognormal distributions with means and , we have also considered the problem of testing hypotheses and computing confidence intervals for . Note that inference concerning is equivalent to that for the ratio of two lognormal means, namely, . Furthermore, we have also provided a procedure for constructing confidence limits for the difference between the lognormal means, namely, . In this article, we have first developed exact confidence intervals and tests for a single lognormal mean using the ideas of generalized p-values and generalized confidence intervals. The resulting procedures are easy to compute and are applicable to small samples. We have also developed similar procedures for obtaining confidence intervals and tests for the ratio (or the difference) of two lognormal means. For each of the problems that we have considered, we have also explained the required computational procedures. A major motivation for the present work is the need for easily computable tests and confidence regions for the lognormal mean, as required in the analysis of occupational exposure data. Our work appears to be the first attempt to obtain small sample inference for the latter problem. We have also compared our test to a large sample test. The conclusion is that the large sample test is too conservative or too liberal, even for large samples, whereas the test based on the generalized p-value controls type I error quite satisfactorily. The large sample test can also be biased, i.e., its power can fall below type I error probability.