Future historians may well label this ''the Age of Statistics.'' Harper's magazine drove the point home when, in March, editor Lewis Lapham added a page headed ''Index.'' The 40 items under that heading come with no comment - just raw data listed menacingly on a single page. Here you can learn that American households watched 213 billion hours of television in 1983; that 1 ,434,668 US documents are currently classified ''Top Secret''; that Carnegie Hall stagehands earn an average of $90,000 a year.
Harper's, of course, didn't inaugurate this fascination with numerical measurement. These days, it is simply a fact of public discourse. One can even assemble data on the data: A cover-to-cover reading of the text (not the charts and tables) of a recent issue of Newsweek, for example, turned up 242 separate bits of statistical information. The word ''percent'' or ''percentage'' appeared 55 times - and was implied (''two-fifths,'' ''one in a hundred,'' and so forth) 20 more times.
Given our numerical fascination, then, it's not surprising that America's statistical granddaddy - the United States Census - should be receiving some attention even in the mid-decade years. The next Census Day is April 1, 1990. Yet this year and next, in some 70 ''Local Public Meetings'' nationwide, the Census Bureau is holding hearings to elicit recommendations on the questions it will ask.
That's all to the good. The issues facing census planners range from the technical (securing accurate mailing lists and good maps) to the political - whether, for example, census-takers should ask the respondent's Spanish/Hispanic origin in addition to his or her race (they do), or ask the respondent's religion (they don't). Peter A. Bounpane, the Census Bureau's chief planner for the 1990 census, notes that only such information as is required by federal law can be gathered. He observes, too, that the questions must be brief and simple: Too much complexity or an appearance of too much snoopi-ness can lead to too many forms being filed in waste-baskets. Even within those constraints, however, there is plenty of flexibility - and plenty of food for thought:
* Place of residence. It sounds simple: Just tell the bureau where you live. But what if you're a ''snowbird'' - maintaining two homes, spending half the year in the North and fleeing south for the winter? The bureau intentionally avoids the term ''legal residence.'' ''We weren't sure we could define it,'' says Mr. Bounpane, in tacit rejection of the legal tangles created by the Internal Revenue Service. Instead, respondents decide which residence to claim. The results, predictably, are a constant concern to states like Florida, which may get undercounted.
* Education. The question of residence for university students has already been settled: They're counted at their dormitory. Other questions remain, however. The 1980 census asked about attendance at a ''regular school or college.'' Given the importance of vocational programs, nondegree courses, and company-sponsored training, many observers wonder whether this is a broad enough definition of education.
* Housing. The census inquires into such things as cost of utilities (used for energy studies) and access to a separate kitchen (used to determine whether the respondent's rooms constitute a separate housing unit). Both raise reporting problems. Mr. Buonpane says that respondents tend to remember the highest recent utility bill - and treat that as an average. And Harvard's William Apgar, who researches housing issues, notes that the respondent's mood can affect answers about such things as ''mother-in-law'' apartments. ''If you're mad at her,'' he notes, ''a hot plate counts as a housing unit.''
That the planners are worried about such fine points suggests the efforts they take to ensure the accuracy of the data. Thrown-away forms, overly complex questions, selective memories, and shifting moods all distort the picture. And irrelevant or loaded questions waste an opportunity to discover really useful information.
The bureau is to be commended for its efforts: If so many statistics are going to be floating around - in our periodicals and on TV, often without any reference to a context - it is imperative that they be carefully gathered. At bottom, however, there are questions behind the questions - ones which, in this Age of Statistics, the nation needs to ask continually. To what extent has the nation benefited from all these data? Because we can count more things, are we necessarily doing better? Or are we knowing more and understanding less?
''A little learning,'' wrote Alexander Pope, emphasizing little, ''is a dangerous thing.'' Pope was not urging ignorance, just cautioning against superficiality. The danger of this fascination with statistics is that it, too, can lead to superficiality - that, content to report the numbers, we stop short of thinking the thoughts. At best, statistics are diagnostic tools. At worst, they are substitutes for profundity. Jefferson, after all, had no census data when he wrote the Declaration of Independence. And de Tocqueville, researching ''Democracy in America,'' never once polled the public pulse.