JUST ADD'

WATER AND

PLUG INz

INSTANT

MOBILE

LABWASHER

Patent Pending

Sets up in seconds. No costly installa-tion charges. And no inconvenience.Quick-disconnect coupling includesuniversal adapter for attachment toany water tap. Smooth-working ballbearing casters lock for stable opera-tion. And a three-wire cord-overeight feet long-makes direct con-nection to distant outlets possible.Operates on 115 Volts-60 cycles.The Labwasher handles almost all

lab glassware. Cleans and dries itthoroughly and efficiently. With 50%less breakage than handwashing.Pays for itself in man-hours saved.

Request Bulletin No. IR 721 formore information.

Find out more. Write to:THE

It J BBEECo.

Department IR721 * 2310 Superior Ave.Cleveland, Ohio

The latter are the recognized authori-ties in their fields. They know becausethey do. And others recognize that theydo. I don't think the dichotomies uni-versity-industry or big-little (science)have much to do with the distinctionbetween basic and applied science (lookat the author bylines in the journals).Decisions based on this distinctionshould be made by. or at least in con-sultation with, the individual recognizedresearchers.

A. W. K. METZNERBrookhaven National Laboratory,Upton, Long Island, New York 11973

. . While exploring the quicksandbetween particle physics and buildingbetter mousetraps, the classical bound-ries of research and science are leftunperturbed. Science is that on whichresearch is done; science and researchinvolve things hard enough to stubone's toe on . . . a Spinco, a micro-scope, or a mousetrap. To me, researchis the Almighty Scientific Method inaction, and if this is true, objectiveevaluation of a teaching program or asystems approach to delivery of healthcare are as much research and as muchscience as a frontal attack on the mys-tery of the gene. Classical research andscience they are not. But the more thatresearch, education, and service are con-sidered as functional aspects of thesame animal, the easier it becomes tomeasure the dollar value, or applica-bility, of the total effort. At one ex-treme, a study may involve no moreservice than providing jobs and no moreeducation than how to operate an elec-tron microscope. At the other, wherestudents are involved in research on aproblem of local service, the learning,discovering, and helping have a positivefeedback relationship toward each oth-er. The value of the system includesthe people in it....

DAVID BRYSONAssociation of A mnerican MedicalColleges, 1501 New HampshireAvenute, NW, Washington, D. C. 20036

Statistical Randomization in theBehavioral Sciences

In his letter (24 Feb.) Stanley takesissue with my protest (Letters, 21 Oct.)against the invalid uses of inferentialstatistics in studies employing (i) non-random samples and (ii) nonrandom as-signment of experimental treatments tosample units. According to Stanley,

314

"hypothesis testing in studies involvingnonrandomly chosen 'grab groups' isfeasible. If it were not, the results ofmany comparative experiments in thebehavioral sciences would be difficultto interpret statistically, because anydifferences whatsoever might be attrib-uted to chance fluctuations."The invalid use of inferential statis-

tics is feasible, possible, and, indeed,popular in the behavioral sciences. Ihasten to agree with the consequenceStanley points out: many studies are"difficult to interpret statistically." Butthis is because these "quasi-experi-ments" to use Stanley's term ("pseudo-experiments" might be less misleading)violate the assumptions of inference.Descriptive statistics are perfectly ap-propriate in these studies, but infer-ence to other populations is a mat-ter of opinion or authority rather thanstatistics.

In the field of educational researchto which Stanley addresses his re-marks, these violations have led toseemingly endless series of equivocal,nonreplicable studies. Important ques-tions such as the effect of class sizeand teaching methods on student learn-ing remain unanswered. Instead oftightly controlled, randomized experi-ments, investigators continue to pursuestatistical significance with ever largersamples in quasi-experiments. In addi-tion, the generality of other nonran-dom behavioral science research forhuman affairs is open to question, es-pecially if the rest of us do not behavelike white rats, pigeons, college stu-dents, pig-tailed macaques, or Maoritribesmen.

HERBERT J. WALBERGGraduate School of Education,Harvard University,Cambridge, Massachutsetts 02138

Growth Incentives in the"Have-Not" Nations

In his address delivered at the De-cember Washington meeting of theAAAS ("The ever widening gap," Sci-ence, 24 Feb., p. 959), Blackett re-iterated the prevailing and plausibleopinion that the population problem inthe poor countries in the South is dueto the export of modern medicine bythe rich countries in the North, whichhas resulted in health, too much andtoo soon, rather than wealth.On the basis of simple arithmetic,

it may seem plausible to construe a

SCIENCE, VOL. 156

U~~~S

$3.00We say "only" because that's half the cost of any otherreusable pipet.Figure it out.A COREX pipet, made in our exclusive chemicaltempering process, lasts in your lab a full 6 timeslonger than any other pipet.Yet, the price of a COREX pipet is only 3 timesthat of any other pipet.So you save half with this new standard.

The new standards come from RNINGLABORATORY PRODUCTS

reductiodh in mortality as the cause'Ofthe population problem, if the prob-lem is viewed simply in terms of therate of population growth. But certain-ly the solution of the problem- is notsimply the achievement or maintenanceof a low growth rate, which could bethe result of a balance either betweenliigh or low birth rates and death rates.All humanitarian considerations aside,only low death rates matched by lowbirth rates will maximize the returnsfrom investments in human resourcesand minimize the ratio of dependentsto producers.The beneficial effects of health serv-

ices are still achieved in synergism withother improvements in the levels ofliving (1). In a reciprocal cause andeffect relationship, mortality declineswith economic development. The ex-tension of health services is a neces-sary but insufficient condition forhealth, which is a necessary althoughinsufficient condition for productivity.Reductions in mortality have alwayspreceded reductions in fertility. Recentevidence confirms that a reduction inmortality is a necessary, although in-sufficient, condition for a reduction infertility (1). As mortality declinesand the population growth rate in-creases, there also develops a trendtoward restraining fertility. Of course,the desired demographic changes areno more automatic than the desiredeconomic changes. The use of the mostefficient, effective, and acceptable meth-ods of family planning would seemto be the most appropriate responseto the spontaneous motivation to limitfamily size.

HARALD FREDERIKSENHealth Service, Agency forInternational Development,Washington, D.C. 20523

References

1. H. Frederiksen, Pub. Health Rept. 81, 715(1966); D. M. Heer, Population ControlSymposium, 133rd AAAS annual meeting,Washington, D.C., Dec. 1966.

I was present when Blackett pre-sented his address. Reading his articlein Science has not changed my first im-pressions. His analysis and statistics onthe low rate of increase of the grossnational product in such countries asIndia, while impressive, end with theusual conclusion that the solution liesin doling out more aid from the"have" to the "have-not" nations.

Blackett's analysis omits the humanfactors that have much to do with the

316

productivity of any people. Certainlythe most important and enduring fac-tor affecting the gross national prod-uct is that people work harder andproduce more when they are assuredprofit and personal gain in proportionto effort and to innovation. This iseven more true in the "poor coun-tries than in the "rich" nations andalso depends on whether one is buck-ing government planning in the UnitedStates, Great Britain, or India. Butthe problem becomes especially severewhen a government's greatest fear isthat some industrialists may becomerich, which seems to be the case inIndia and possibly also in Great Brit-ain. Although the United States govern-ment adds many controls, it fortunate-ly does not lose sight of the fact thatlarger personal and corporate profitsare the golden eggs that produce a

larger GNP, increased taxable income,and a higher standard of living.

I have attended a week-long confer-ence on problems of undevelopedcountries where "profit," "industrialcompetition," and "free enterprise"were forgotten phrases and nearlyevery analysis ended with demandsfor more doling of outside aid. Anal-yses that disregard the human drivesthat produced the industrial revolutionare not likely to point the way to anybetter solutions. While more R&D andtechnology are needed, they must bearfruit within human organizations. Asan example, because of the absenceof individual motivation, even thoughthe Soviet Union stands in the frontlines of science and technology it stillcannot comfortably produce foodstuffor other commodities in proportion toher need or great technological capac-ity.

Let's continue to help each otherand to extol the virtues of R&D andof technological information; but let'snot disregard the human factors thatare needed to turn these into a largerGNP. Let the governments of the"poor" countries reduce their bureau-cratic "planning" by which nearlyeveryone remains poor, and insteadaddress their attention to helping manythousands more to become "rich," tax-ably rich, through production of goodsand through competition that dependson invention and innovation ratherthan on preferred position with thegovernment.

V. L. PARSEGIANRensselaer Polytechnic Institute,Troy, Nlew York

SCIENCE, VOL. 156

-

Walntto

Beckmanhas combinedUVand IR

into one

compactpackage...

like this, for only $67100

Now for the first time you can getUV and IR analysis capability inone simple compact instrumentpackage-with a convenient switchto employ the integral recorder foreither mode. Only Beckman offers it.And you can get this package-a

complete UV-IR spectroscopy labor-atory-for $3500 less than you canget a combination of instruments todo the same job from ANY othersource.

We simply combined Beckman'sDBa spectrophotometer and theMicrospecTM IR spectrophotometerin this new way, and added the oiled-walnut cabinet. Both instrumentsare proven for the educational, ex-

perimental, or commercial labora-tory. Proven for versatility. Provenfor reliability and accuracy. And weincluded a book of practical UV-IRexperiments-plus most of the re-

quired sample handling equipment.

You can't find an instrument pack-age (or price) to match this any-where else. So contact your nearbyBeckman Sales and Service Office,or send for Data File No. 34, andlearn all about it now.

INSTRUMENTS, INC.

SCIENTIFIC AND PROCESSINSTRUMENTS DIVISION

FULLERTON, CALIFORNIA * 92634

INTERNATIONAL SUBSIDIARIES: GENEVA; MUNICH; GLENROTHES, SCOTLAND;TOKYO; PARIS; CAPETOWN. LONDON; MEXICO CITY

32 SCEN VOL 156 ----

148k,... ..' li.,.% 1..

SCI1ENCE. VOL. 156320(