We present the results of a 10 ks XMM-Newton observation of NGC 3998 , a “ type-I ” LINER galaxy ( i.e. , with significant broad H \alpha 
emission ) .
Our goal is to test the extent to which radiatively-inefficient accretion flow ( RIAF ) models and/or scaled-down AGN models are consistent with the observed properties of NGC 3998 .
A power-law fit to the XMM-Newton spectra results in a power-law slope of \Gamma = 1.9 and 2-10 keV flux of 1.1 \times 10 ^ { -11 } 
erg cm ^ { -2 } s ^ { -1 } , in excellent agreement with previous hard X-ray observations .
The OM UV flux at 2120Å appears to be marginally resolved , with \sim 50 \% of the flux extended beyond 2 ” .
The nuclear component of the 2120Å flux is consistent with an extrapolation of the X-ray power-law , although \sim 50 \% of the flux may be absorbed .
The OM U flux lies significantly above the X-ray power-law extrapolation and contains a significant contribution from extragalactic emission .
The upper-limit for narrow Fe-K emission derived from the XMM-Newton 
spectra is 33 eV ( for \Delta \chi ^ { 2 } = 2.7 ) .
The upper-limit for narrow Fe-K emission derived from a combined fit of the XMM-Newton and BeppoSAX 
spectra is 25 eV , which is one of the strictest limits to date for any AGN .
This significantly rules out Fe-K emission as is expected to be observed in typical Seyfert 1 galaxies .
The X-ray flux of NGC 3998 has not been observed to vary significantly ( at > 30 \% level ) within the X-ray observations , and only between observations at a level of \sim 50 \% , which is also in contrast to typical Seyfert 1 galaxies .
The lack of any 
reflection features suggests that any optically-thick , geometrically-thin accretion disk must be truncated , probably at a radius of order 100-300 ( in Schwarzschild units ) .
RIAF models fit the UV to X-ray spectral energy distribution of NGC 3998 reasonably well .
In these models the mid-IR flux also constrains the emission from any outer thin disk component that might be present .
The UV to X-ray SED is also consistent with a Comptonized thin disk with a very low accretion rate ( \dot { M } < 10 ^ { -5 } \dot { M } _ { Edd } ) , in which case the lack of Fe-K emission may be due to an ionized accretion disk .
Accretion models in general do not account for the observed radio flux of NGC 3998 , and the radio flux may be due to a jet .
Recent jet models may also be consistent with the nuclear fluxes of NGC 3998 in general , including the X-ray , optical/UV and mid-IR bands .
The ( ground-based ) near-IR to optical photometric data for the nuclear region of NGC 3998 contains large contributions from extra-nuclear emission .
We also derive nuclear fluxes using archival HST WFPC2 data , resulting in meaningful constraints to the nuclear SED of NGC 3998 in the optical band .
We discuss a possible OM U band and USNO-B detection of the NGC 3998 ULX .